Intelligent vehicle network interaction control system and method for large-scale electric vehicles

Abstract

The application discloses a large-scale electric vehicle-oriented intelligent vehicle-grid interaction control system and method, relates to the technical field of electric vehicle and grid interaction, and comprises the following steps: constructing a software-defined charging network architecture, establishing a control plane and a data plane, the control plane comprising a power distribution cloud master station and a district intelligent fusion terminal, and the data plane comprising V2G charging piles and electric vehicles; the control plane and physical devices are decoupled, the control plane issues charging and discharging scheduling strategies and power instructions to the data plane, and the data plane feeds back operation data to the control plane. The application constructs a software-defined vehicle-grid interaction architecture, integrates hybrid prediction, distributed decision-making, multi-target optimization, ISO15118 bidirectional communication and block chain settlement, and realizes efficient scheduling of large-scale V2G resources, safe communication and trusted transaction.

Description

Technical Field

[0001] This invention relates to the field of electric vehicle and power grid interaction technology, and more specifically to an intelligent vehicle-grid interaction control system and method for large-scale electric vehicles. Background Technology

[0002] With the rapid growth of electric vehicle ownership, the scale of electric vehicles connected to the power grid continues to expand. Large-scale, unregulated charging of electric vehicles will significantly impact grid operation, exacerbating peak-valley differences and increasing the risk of overload on distribution equipment, thus affecting the safety and stability of grid operation. Meanwhile, electric vehicles, as a distributed energy storage resource, can achieve bidirectional energy exchange between vehicles and the grid through V2G (Vehicle-to-Grid) technology. While meeting user travel needs, they can also transmit energy from vehicle batteries back to the grid, participating in peak shaving, frequency regulation, and demand response, thereby improving the economy and reliability of grid operation. At the communication level, the ISO 15118 standard supports seamless certification, real-time electricity price integration, and reverse charging functionality, complying with the relevant requirements of the Alternative Fuels Infrastructure Regulations (AFIR), providing a standardized communication foundation for large-scale vehicle-to-grid interaction.

[0003] However, existing V2G control strategies mainly adopt a centralized control architecture, which suffers from a single point of failure in the centralized controller, leading to the paralysis of the entire dispatching system. Furthermore, centralized control response exhibits lag, making it difficult to adapt to millisecond-level dynamic changes in grid load. The control strategy lacks flexibility, failing to implement personalized dispatching based on different users' travel preferences and battery health status. As the number of connected vehicles continues to expand, the computational burden on the centralized controller increases exponentially, resulting in poor scalability. In addition, regarding communication protocols, the basic AC charging protocol conforming to the IEC 61851-1 standard lacks plug-and-charge (PnC) and bidirectional power transfer (BPT) functions, making it difficult to meet the needs of high-level vehicle-to-grid interaction. The existing technology's integration with trusted settlement mechanisms such as blockchain is insufficient, leading to high charging and discharging transaction settlement costs, low transparency, and an imperfect trust mechanism. Therefore, there is an urgent need for an intelligent vehicle-to-grid interaction control method and system that can support large-scale electric vehicle access, possess distributed decision-making capabilities, and balance grid benefits with user satisfaction.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent vehicle-to-everything (V2X) interactive control system and method for large-scale electric vehicles, in order to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an intelligent vehicle-to-grid (V2G) interactive control method for large-scale electric vehicles, comprising the following steps: A software-defined charging network architecture is constructed, establishing a control plane and a data plane. The control plane includes the distribution cloud master station and the intelligent converged terminal of the transformer area, while the data plane includes V2G charging piles and electric vehicles. This achieves decoupling between the control plane and physical devices, and the control plane issues charging and discharging scheduling strategies and power commands to the data plane, while the data plane feeds back operating data to the control plane. Real-time data collection of vehicle status, charging pile operation, and power grid load is used. A hybrid prediction model combining OPTICS density clustering and EEMD-CLSTM-GWO is employed to predict future power grid load demand and electric vehicle charging and discharging potential, yielding prediction results. The prediction results and real-time vehicle status data are obtained, and the local decision-maker in the V2G charging pile calculates the optimal charging and discharging power command based on a multi-objective optimization model, and generates the corresponding charging and discharging control strategy. Based on the charging and discharging power command, the bidirectional power conversion module is controlled to perform charging or discharging operations, and bidirectional data interaction between the V2G charging pile and the electric vehicle is completed based on the ISO 15118-20 communication protocol. The charging and discharging process data is written into the blockchain network, and the smart contracts deployed in the blockchain network record the charging and discharging data, automatically calculate the compensation amount based on the actual charging and discharging amount, and complete the charging and discharging transaction settlement.

[0007] The hybrid prediction model includes the OPTICS density clustering model, the EEMD decomposition model, the CLSTM prediction model, and the GWO optimization model. The OPTICS density clustering model is used to cluster and identify user groups with different usage patterns based on user vehicle operation characteristics. The EEMD decomposition model is used to perform multi-scale decomposition of non-stationary power grid load signals. The CLSTM prediction model is used to extract the time-series variation characteristics corresponding to each decomposition component and complete the load trend prediction. The GWO optimization model is used to globally optimize the model parameters of the CLSTM prediction model.

[0008] The local decision-makers in V2G charging piles adopt a distributed decision-making mechanism. Each local decision-maker independently calculates the charging and discharging power commands based on the corresponding electric vehicle status data and prediction results. The distributed decision-making mechanism does not rely on the centralized controller to complete single-node decisions. When any local decision-maker malfunctions, the remaining local decision-makers continue to operate independently and maintain the charging and discharging control of the corresponding electric vehicle according to the preset safety strategy.

[0009] When the bidirectional power conversion module is in the discharge working state, it adjusts the discharge power output from the vehicle to the grid according to the grid frequency deviation value, so that the electric vehicle participates in the primary frequency regulation of the grid. The response time of the primary frequency regulation is less than or equal to 100ms.

[0010] The V2G authorized smart contract automatically calculates the compensation amount based on the actual charging and discharging volume of the electric vehicle. The compensation amount is determined based on the discharge volume, the discharge compensation unit price, the charging volume, and the charging unit price. The calculation relationship is as follows: in, Indicates the amount of discharge. This indicates the unit price for discharge compensation. Indicates the charging capacity. This indicates the unit price for charging.

[0011] V2G authorized smart contracts are used to automate the management of charging and discharging authorization, charging and discharging execution, discharge quantity calculation and fee settlement processes. They also store the scheduling instructions, execution data and settlement data in the blockchain distributed ledger during the charging and discharging interaction process, forming a trusted interaction mechanism based on the entire process of instruction-execution-settlement.

[0012] The intelligent vehicle-to-grid (V2G) interaction control system for large-scale electric vehicles includes a power distribution cloud master station, intelligent fusion terminals for distribution areas, V2G interaction protocol conversion devices, V2G charging piles, and a blockchain network. The distribution cloud master station is deployed in the power grid dispatch center to receive power grid dispatching requests, perform hierarchical aggregation of dispatchable capacity, and generate regional V2G dispatching instructions. At least one smart converged terminal for the distribution area is deployed on the distribution area side corresponding to the distribution transformer and is connected to the distribution cloud master station. It is used to receive regional V2G dispatch instructions, aggregate the operation data of charging piles in the distribution area, distribute regional V2G dispatch instructions to the corresponding V2G charging piles, and execute local dispatch decisions. At least one vehicle-to-grid (V2G) communication protocol conversion device is connected between the smart fusion terminal in the distribution area and the V2G charging pile to realize protocol conversion and communication adaptation between different communication protocols. At least one V2G charging pile is connected to the vehicle-to-grid interaction protocol conversion device. The V2G charging pile includes a local decision-maker, a bidirectional power conversion module and a communication module. The local decision-maker has a built-in multi-objective optimization model, which is used to calculate the charging and discharging power command based on real-time operating data. The bidirectional power conversion module executes charging or discharging operations according to the charging and discharging power command. The communication module is used to complete the data interaction between the V2G charging pile and the electric vehicle and the intelligent integrated terminal of the distribution area. The blockchain network communicates with V2G charging piles and smart integrated terminals in the distribution area. The blockchain network is equipped with V2G authorized smart contracts to record charging and discharging data and to execute charging and discharging transaction settlements based on the charging and discharging data.

[0013] The local decision-maker incorporates a multi-objective optimization model. This model uses grid-side operating costs, battery cycle aging costs, and user satisfaction as optimization objectives. The objective function is expressed as follows: in, This represents the overall optimization target value; This represents the grid-side operating cost, used to characterize the net cost between the electricity purchase cost and the revenue from peak shaving and frequency regulation during the charging and discharging process; This represents the cost of battery cycle aging, used to characterize the lifespan loss of a vehicle battery due to charging and discharging. It represents user satisfaction and is used to characterize the extent to which the current state of battery charge of the vehicle meets the user's travel needs; , , These represent the weighting coefficients for grid operation costs, battery cycle aging costs, and user satisfaction, respectively. The weighting coefficients are configured based on grid dispatch requirements, battery type, and user operation mode.

[0014] The battery cycle aging cost satisfies the following relationship: in, This represents the battery aging factor, which is determined based on the battery's state of charge (SOC), charge / discharge rate (C-rate), and battery temperature. Indicates charging power or discharging power; Indicates the duration of charging and discharging; User satisfaction satisfies the following relationship: in, This indicates the user-defined threshold for the state of charge for travel protection. Indicates the vehicle's current actual state of charge; when At that time, the user satisfaction value is 1.

[0015] The local decision-maker uses model predictive control or dynamic programming algorithms to solve the multi-objective optimization model and generates charging and discharging power commands based on the prediction results, vehicle status data, and grid load data. When using model predictive control, the prediction time domain is 15-30 minutes and the control time domain is 5-15 minutes. The communication module supports the ISO15118-20 communication protocol stack. Based on the ISO15118-20 communication protocol, the communication module completes bidirectional communication between the V2G charging pile and the electric vehicle. It supports bidirectional charging modes such as vehicle-to-grid V2G, vehicle-to-home V2H, and vehicle-to-building V2B. The communication process includes TLS1.3 secure transmission layer encryption, plug-and-charge authentication based on digital certificates, and bidirectional power transmission parameter negotiation, enabling the V2G charging pile to switch between charging and discharging modes according to charging and discharging power commands.

[0016] The technical effects and advantages provided by the present invention in the above technical solution are as follows: This invention employs a software-defined charging network architecture, decoupling the control plane from physical devices. This separates the scheduling decision logic for charging resources from the underlying hardware, allowing control strategies to be optimized and updated without relying on specific device structures. The distribution cloud master station is responsible for global scheduling management, the intelligent converged terminals in the distribution areas handle edge coordination, and the V2G charging piles are responsible for specific charging and discharging execution, forming a hierarchical collaborative control system. As the number of connected vehicles continues to increase, resource allocation can be completed simply by updating the scheduling strategy of the control plane, without adjusting the underlying device structure. This improves the flexibility and scalability of the entire vehicle-to-grid interaction control system, supporting a single aggregator to manage more than 100,000 electric vehicles simultaneously participating in charging and discharging scheduling, thus improving scheduling efficiency and resource management capabilities in large-scale electric vehicle access scenarios.

[0017] This invention employs a hybrid prediction model comprised of OPTICS density clustering, EEMD, CLSTM, and GWO, forming a complete prediction system covering user behavior identification, load signal processing, time-series prediction, and model optimization. OPTICS density clustering identifies user groups with different vehicle usage patterns, providing user profiles for differentiated scheduling; EEMD decomposes non-stationary grid load signals, improving the ability to process complex load data; CLSTM extracts the time-series variation characteristics of each decomposed component, taking into account both short-term load fluctuations and long-term trend predictions; and GWO further optimizes the global hyperparameters of CLSTM, improving model prediction accuracy. The multiple algorithms form a collaborative prediction mechanism, making the prediction of future grid load demand, electric vehicle charging and discharging potential, and renewable energy power generation more accurate, with prediction errors controlled within 8%, providing reliable data support for subsequent optimized scheduling.

[0018] This invention employs a distributed decision-making mechanism, delegating charging and discharging decision-making power to local decision-makers in each V2G charging pile. Each local decision-maker independently calculates charging and discharging power based on prediction results and the vehicle's real-time status, eliminating the reliance on a centralized controller to handle all scheduling tasks and effectively removing single-point bottlenecks in centralized control architectures. Each local decision-maker can perform synchronous optimization calculations, resulting in faster scheduling response and a system response speed improvement of over 50%. When any local decision-maker fails, the remaining local decision-makers continue to operate normally, with the faulty node automatically entering a safe mode without affecting the overall scheduling function. This achieves node fault isolation and local autonomous control, further enhancing the operational stability, reliability, and continuous service capability of the entire vehicle-to-grid (V2G) interactive control system.

[0019] This invention constructs a multi-objective optimization model that simultaneously considers three optimization objectives during the charging and discharging control process: grid efficiency, battery life, and user experience. Specifically, the Cgrid term comprehensively evaluates electricity purchase costs and peak-shaving and frequency regulation benefits to achieve peak shaving and valley filling; the Cbattery term comprehensively considers battery state of charge, charge / discharge rate, and battery temperature to reduce the impact of frequent charging and discharging on battery life; and the Ruser term ensures that the vehicle's actual state of charge meets users' travel needs, improving user satisfaction. These multiple optimization objectives are calculated collaboratively within a unified model, enabling the charging and discharging power to balance grid operation requirements, vehicle operating status, and user needs, achieving a coordinated and unified benefit among the grid, users, and vehicle operators, and improving the overall economic efficiency of the vehicle-grid interaction system.

[0020] This invention is compatible with the ISO 15118-20 communication standard. Without altering the existing interface, it establishes a bidirectional communication mechanism between vehicles and V2G charging stations, supporting V2G, V2H, and V2B bidirectional charging use cases. This enables secure and controllable bidirectional power flow between vehicles, charging stations, and the power grid dispatching system. The communication process employs TLS 1.3 secure transport layer encryption, combined with plug-and-charge authentication based on digital certificates and bidirectional power transmission parameter negotiation, enhancing communication security and device interconnectivity. This allows vehicles to quickly complete identity authentication and bidirectional charging / discharging control. Furthermore, it complies with the EU AFIR regulation's requirement for charging stations to support the ISO 15118-20 standard from January 1, 2027, improving system standard compatibility and international application adaptability.

[0021] This invention deploys V2G authorized smart contracts in a blockchain network to achieve fully automated management of the entire charging and discharging transaction process. The smart contract automatically completes vehicle selection, compensation calculation, and transaction settlement based on actual vehicle charging and discharging data. Simultaneously, it records dispatch instructions, the charging and discharging execution process, and the revenue settlement results to the blockchain distributed ledger, forming an immutable data record. The entire charging and discharging transaction process requires no manual intervention in settlement, improving revenue calculation efficiency and reducing manual processing costs by 70%. Furthermore, relying on the blockchain distributed ledger ensures transparent and fully traceable transaction data, constructing a comprehensive trust mechanism covering the entire process from instruction to execution to settlement, further enhancing user trust and transaction credibility.

[0022] This invention focuses on large-scale electric vehicle (EV) vehicle-to-grid (V2G) interaction scenarios, integrating several key technologies such as software-defined charging networks, hybrid prediction models, distributed decision-making, multi-objective optimization, ISO 15118-20 bidirectional communication, and blockchain smart contracts to form a complete intelligent management and control system for large-scale V2G resources. The system can dynamically adjust charging and discharging strategies based on future grid load changes and vehicle operating status, improving the flexibility of EV cluster scheduling and resource utilization efficiency, while also enhancing load forecasting accuracy, charging and discharging control precision, and overall system economy. The entire technical solution is applicable to large-scale EV access scenarios at the city level, providing reliable support for large-scale V2G resources to participate in grid peak shaving, frequency regulation, and demand response, and possesses significant commercialization value and application prospects. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0024] Figure 1 This is a flowchart of the intelligent vehicle-to-grid (V2G) interactive control method for large-scale electric vehicles according to the present invention.

[0025] Figure 2 This is a schematic diagram of the modules of the intelligent vehicle-to-grid interactive control system for large-scale electric vehicles according to the present invention. Detailed Implementation

[0026] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0027] This invention provides, for example Figure 1 The intelligent vehicle-to-grid (V2G) interaction control method for large-scale electric vehicles, as shown, has the following specific steps: A software-defined charging network architecture is constructed, comprising a control plane and a data plane. The control plane includes the distribution cloud master station and intelligent converged terminals in the distribution areas, while the data plane includes V2G charging piles and vehicles. The control plane issues charging and discharging scheduling strategies and power commands to the data plane via a southbound interface, while the data plane reports real-time operational data to the control plane via a northbound interface. This software-defined charging network architecture separates the scheduling decision logic of charging resources from the underlying physical devices, allowing the upper-layer scheduling strategies to be optimized and updated independently of specific hardware, thereby improving the system's flexibility and scalability.

[0028] The specific implementation process is as follows: First, a logical mapping relationship is established between the control plane and the data plane. The distribution cloud master station, as the upper-level scheduling node of the control plane, is responsible for receiving regional scheduling requests from the power grid dispatch center and generating corresponding charging and discharging scheduling strategies based on the available scheduling capacity within the region. The intelligent fusion terminal of the distribution area, as the edge scheduling node of the control plane, maintains a communication connection with the distribution cloud master station and manages the V2G charging piles within its respective distribution area in a unified manner, forming a control system that combines regional and distribution area levels. The data plane consists of multiple V2G charging piles and connected vehicles. Each V2G charging pile establishes a communication connection with its corresponding vehicle to achieve real-time interaction of vehicle operating status, charging and discharging status, and equipment operating status, enabling the control plane to continuously obtain operating information from the underlying equipment.

[0029] After the control plane generates the scheduling strategy, it sends the charging / discharging scheduling strategy and power commands to the data plane via the southbound interface. The southbound interface serves as the command transmission channel between the control plane and the data plane, progressively transmitting the regional-level scheduling target to the corresponding distribution areas. The distribution area's intelligent fusion terminal then distributes these targets to the corresponding V2G charging piles, enabling each V2G charging pile to execute the corresponding charging or discharging mode according to the received scheduling strategy. The charging / discharging scheduling strategy includes the target charging / discharging direction, target power information, and scheduling execution cycle. The power commands constrain the output power of each V2G charging pile, ensuring that all charging devices operate collaboratively according to a unified scheduling target.

[0030] During the charging and discharging control process of each V2G charging pile, real-time operating data of the corresponding vehicle and equipment is continuously collected and fed back to the control plane via the northbound interface. The northbound interface serves as a data upload channel, uploading the real-time operating data collected from the data plane to the intelligent fusion terminal of the distribution area. The intelligent fusion terminal then aggregates the data and sends it to the distribution cloud master station, enabling the control plane to continuously perceive the operating status of the data plane. Real-time operating data includes the vehicle's current state of charge, battery capacity, estimated downtime, user travel plan, battery health status, current output power of the charging pile, operating temperature, cumulative charging and discharging amount, real-time load rate of the distribution area transformer, feeder current, feeder voltage, and regional power grid frequency deviation information. All types of operating data collectively form the data foundation for subsequent scheduling decisions.

[0031] As real-time operational data is continuously fed back to the control plane, the distribution cloud master station, combined with data uploaded by the smart fusion terminals in the distribution areas, comprehensively analyzes the operational status of each distribution area within the region and dynamically adjusts the charging and discharging scheduling strategy based on changes in the dispatchable capacity of each distribution area. When the regional load changes, the distribution cloud master station regenerates the regional-level V2G scheduling command and sends it again to the data plane via the southbound interface; the smart fusion terminals in the distribution areas redistribute the command according to the new scheduling target, enabling each V2G charging pile to synchronously update its charging and discharging control status, thus ensuring that the control strategy remains consistent with the grid's operational status.

[0032] When communication between the distribution cloud master station and the smart fusion terminal of the distribution area is normal, the regional dispatch is uniformly coordinated by the distribution cloud master station, and each distribution area executes charging and discharging control according to the unified dispatch target. When communication between the distribution cloud master station and the smart fusion terminal of the distribution area is interrupted, the smart fusion terminal of the distribution area continues to maintain local dispatch within the distribution area based on the locally stored operation data and dispatch information, so that the V2G charging piles under its jurisdiction continue to maintain normal operation. After communication is restored, the data generated during local operation is synchronously uploaded to the distribution cloud master station, so that the control plane can re-complete the regional operation status update and ensure that the entire charging and discharging dispatch process remains continuous.

[0033] Because the software-defined charging network architecture separates the scheduling decision-making logic of charging resources from the underlying physical devices, the optimization, adjustment, and updating of scheduling strategies are all concentrated on the control plane, while the data plane continues to maintain its original equipment structure to complete charging and discharging execution. Scheduling strategy switching can be completed without adjusting the underlying hardware configuration. As the number of connected V2G charging piles and electric vehicles continues to increase, unified management of vehicle-to-grid interaction resources of different scales can be achieved simply by adjusting the scheduling strategy of the control plane. The control plane continuously updates the scheduling strategy based on real-time operating data, and the data plane completes charging and discharging control based on the latest scheduling strategy. The two maintain an independent yet real-time collaborative operating relationship, thereby improving the flexibility, scalability, and scheduling efficiency of the entire vehicle-to-grid interaction control system when facing large-scale electric vehicle access scenarios.

[0034] Real-time collection of multi-dimensional data, including vehicle status data, charging pile operation data, and power grid load data. Vehicle status data includes battery state of charge (SOC), battery capacity, estimated downtime, user travel plans, and battery state of health (SOH). Charging pile operation data includes charging pile identification, current output power, operating temperature, and cumulative charge and discharge amount. Power grid load data includes real-time load rate of transformer substations, feeder current, feeder voltage, and regional power grid frequency deviation. Based on the collected data, a hybrid prediction model combining OPTICS density clustering and EEMD-CLSTM-GWO is used to predict future grid load demand, renewable energy generation capacity, and the charging and discharging potential of electric vehicle clusters. Specifically, OPTICS (Ordering Points To Identify the Clustering Structure) density clustering algorithm is used to cluster and identify user groups with different usage patterns, distinguishing between commuter, business, and leisure users, providing a user profile basis for subsequent differentiated scheduling. EEMD (Ensemble Empirical Mode Decomposition) is used to decompose the non-stationary grid load signal, dividing the original load sequence into multiple intrinsic mode function components and a residual component. CLSTM (Convolutional Long Short-Term Memory) network is used to extract and predict the temporal features of each decomposed component, capturing short-term fluctuations and long-term trends in the load sequence. GWO (Grey Wolf) is used to predict the load sequence. The Optimizer (Grey Wolf Optimization Algorithm) performs global optimization of the hyperparameters of CLSTM, improving prediction accuracy and keeping the prediction error of the hybrid prediction model within 8%.

[0035] The specific implementation process is as follows: the control plane continuously receives operational information from the data plane and establishes a data acquisition mechanism corresponding to the vehicle's operating status, charging facility operating status, and power grid operating status. After a vehicle connects to a V2G charging pile, the current operating parameters of the vehicle are first read to form vehicle status data. Among these, the battery state of charge reflects the vehicle's current remaining power level, the battery capacity characterizes the vehicle's energy storage capacity, the estimated downtime reflects the schedulable duration of the vehicle's participation in vehicle-to-grid interaction, the user's travel plan reflects the vehicle's future departure time and travel demand, and the battery health status reflects the current performance status of the power battery. All these data together form the vehicle-side operating characteristics. At the same time, the V2G charging pile continuously records its own operating status, acquiring the charging pile identifier, current output power, operating temperature, and cumulative charge and discharge amount in real time, so that the charging equipment's operating status is kept synchronized with the vehicle's status. On the distribution side, the real-time load rate of the transformer, feeder current, feeder voltage, and regional power grid frequency deviation are collected simultaneously, so that the power grid operating status can reflect the current load changes in real time. The multi-source data together constitute the data input for subsequent predictive modeling.

[0036] After completing multi-dimensional data collection, various operational data were categorized and organized according to data type, and then uniformly input into the hybrid prediction model. First, the OPTICS density clustering algorithm was used to process the vehicle operation data, using estimated vehicle downtime, user travel plans, and vehicle operation patterns as clustering criteria to identify user groups with different usage patterns, distinguishing between work commuters, business users, and leisure users. Work commuters have relatively fixed vehicle parking times and travel cycles, business users exhibit greater time fluctuations, and leisure users show stronger randomness. These different categories form different user profiles, providing a data foundation for subsequent differentiated scheduling and enabling the prediction model to analyze the operation patterns of different vehicle categories.

[0037] After user clustering is completed, the power grid load data is input into the EEMD processing procedure. Because the power grid load signal exhibits significant non-stationary characteristics, with varying frequency trends across different time periods, an ensemble empirical mode decomposition (EMD) method is employed to decompose the non-stationary power grid load signal. This decomposes the original load sequence into multiple intrinsic mode function (IMF) components and a residual component. Each IMF corresponds to load variation characteristics within a different frequency range, while the residual component reflects the overall load variation trend. After decomposition, the complex load sequence is transformed into multiple data components with different time scale characteristics, providing a more stable data foundation for subsequent time series forecasting and reducing the impact of the original non-stationary signal on the forecast results.

[0038] Subsequently, each intrinsic mode function component and residual component are input into the CLSTM model for prediction processing. The CLSTM utilizes a convolutional structure to extract local variation features between components, and then uses a long short-term memory network to analyze the temporal variation patterns of each component, establishing prediction results for load change processes at different time scales, thereby simultaneously capturing both short-term fluctuations and long-term trends in the load sequence. After each component's prediction is completed, the prediction results are integrated to form a future grid load demand prediction. Simultaneously, combined with vehicle clustering results and charging pile operating status, the charging and discharging potential of future electric vehicle clusters that can participate in dispatch is predicted. Furthermore, the future renewable energy power generation is simultaneously predicted based on the current operating status, ensuring that load demand prediction, renewable energy prediction, and vehicle-grid interaction resource prediction maintain a unified time scale.

[0039] To further improve the accuracy of the prediction model, the Grey Wolf Optimization (GWO) algorithm was introduced to optimize the model parameters after the CLSTM model was established. The GWO algorithm performs global optimization around the hyperparameters of the CLSTM model, continuously adjusting the model parameter configuration based on the prediction error, gradually converging the model to a better parameter combination and improving the accuracy of load prediction results at various time scales. The optimized CLSTM model then performs prediction calculations again, ensuring high consistency between the predicted future grid load demand, renewable energy generation capacity, and the charging and discharging potential of electric vehicle clusters, and feeding the prediction results back into the subsequent charging and discharging scheduling process.

[0040] After the forecast is completed, comprehensive forecast information is generated, including forecasts of future grid load demand, future renewable energy power generation, and the future charging and discharging potential of electric vehicle clusters. This comprehensive forecast information is sent to the local decision-maker as the data basis for optimizing charging and discharging power calculations. Because the hybrid forecast model employs four models working collaboratively—OPTICS density clustering, EEMD signal decomposition, CLSTM time series forecasting, and GWO parameter optimization—each model undertakes user classification, signal processing, trend prediction, and parameter optimization functions, different types of data can be processed according to their corresponding characteristics. Ultimately, the forecast error of the hybrid forecast model is controlled within 8%, providing stable and accurate data support for subsequent charging and discharging scheduling and vehicle-grid interaction control.

[0041] Each V2G charging station incorporates a local decision-maker. This local decision-maker acquires the prediction results from the hybrid prediction model and real-time vehicle status data, and calculates the optimal charging and discharging power based on a multi-objective optimization model. The multi-objective optimization model comprehensively considers three optimization objectives: grid-side operating costs, battery cycle aging costs, and user satisfaction. It generates charging and discharging power commands while meeting both vehicle operation and grid dispatch requirements. Its objective function is expressed as: in, This represents the comprehensive optimization objective function, used to measure the degree of comprehensive optimization of the current charging and discharging scheme; , , The weight coefficient represents the corresponding optimization objective and is used to adjust the degree of influence of different optimization objectives in the comprehensive optimization process. The weight coefficient is dynamically adjusted according to the priority of grid peak-shaving demand, battery type and user selection mode. Indicates grid-side costs; Indicates the cost of battery cycle aging; This represents a user satisfaction constraint.

[0042] The local decision-maker first reads the predicted future grid load demand, the charging and discharging potential of the electric vehicle cluster, and the current vehicle operating status. Based on these input data, it establishes the current charging and discharging control state. Within a unified optimization framework, it simultaneously analyzes the grid operating status, battery life status, and user vehicle demand, ensuring that the three optimization objectives are calculated collaboratively, rather than determining the charging and discharging strategy based on a single indicator. As the vehicle operating status and grid load status continuously change, the local decision-maker synchronously updates the parameters corresponding to the objective function and recalculates the comprehensive objective function value, enabling the charging and discharging power to dynamically adjust to the current operating status.

[0043] Grid-side costs are used to reflect the impact of the charging and discharging process on the economic operation of the power grid, and their expression is as follows: in, This indicates the discharge power output by the vehicle to the power grid; This indicates the charging power output from the power grid to the vehicle; Indicates peak hour electricity price; Indicates off-peak electricity price; This represents the charging electricity price. When the vehicle is in discharge mode, the revenue from peak shaving and frequency regulation is calculated based on the discharge power and the peak-valley electricity price difference; when the vehicle is in charging mode, the corresponding charging cost is calculated based on the charging power and the charging electricity price. The two parts together constitute the grid-side cost, enabling charging and discharging control to take into account the economic operation goals of the power grid.

[0044] Battery cycle aging cost reflects the lifespan loss of a vehicle battery due to charge-discharge cycles, and its expression is as follows: in, This represents the battery aging factor, which is related to the battery's state of charge (SOC), charge / discharge rate (C-rate), and battery temperature. It represents the absolute value of charging or discharging power, used to indicate the actual power involved in the cycle; Indicates the duration of charging and discharging; This represents the time element in the integral calculation process. As the charging and discharging power increases or the charging and discharging duration extends, the battery cycle aging cost changes synchronously. This allows the local decision-maker to consider the vehicle battery life when calculating the charging and discharging power, avoiding the accelerated battery aging caused by long-term high-power charging and discharging.

[0045] User satisfaction reflects the degree to which the vehicle's current battery charge meets the user's travel needs, and its expression is: in, This indicates the user-defined threshold for the state of charge for travel protection. This indicates the vehicle's current actual state of charge. When the vehicle's current actual state of charge reaches or exceeds the user-set travel guarantee state of charge threshold, user satisfaction is calculated based on the relationship between the current state of charge and the travel guarantee state of charge threshold; when the vehicle's current actual state of charge is lower than the user-set travel guarantee state of charge threshold, the user satisfaction value is 0, thus ensuring that charging and discharging control prioritizes meeting the vehicle's normal travel needs.

[0046] After constructing the objective function, the local decision-maker uses either Model Predictive Control (MPC) or Dynamic Programming (DP) to solve the optimization model and generate charging / discharging power commands. When using MPC, the prediction time domain is 15–30 minutes, and the control time domain is 5–15 minutes. The local decision-maker predicts the charging / discharging power based on future load changes and vehicle operating status within the prediction time domain, and then adjusts the power output in real time based on the control time domain. This ensures that the calculated charging / discharging power continuously adapts to changes in grid load and vehicle operating status, improving the dynamic response capability of charging / discharging control. When using DP, the local decision-maker optimizes the solution based on the state change relationship between the current state and the target state to obtain charging / discharging power commands that meet the objective function requirements.

[0047] During the optimization process, multiple operational constraints are simultaneously satisfied. The power constraint satisfies: in, Indicates the first The charging and discharging power corresponding to each V2G charging pile; This indicates the minimum charging and discharging power that the charging station is allowed to output; This indicates the maximum allowable charging and discharging power of the charging pile, ensuring that each V2G charging pile always operates within the allowable power range.

[0048] The battery state of charge constraint is satisfied as follows: in, Indicates the vehicle's current state of charge; This indicates the lower limit of discharge, typically set to 20%. This indicates the charging limit, typically set to 90%. During the vehicle's charging or discharging process, the battery's state of charge remains within the allowable range to ensure normal vehicle operation while also considering battery lifespan.

[0049] The transformer capacity constraint is satisfied as follows: in, This represents the total charging and discharging power of all V2G charging piles within the same area. This indicates the remaining available capacity of the transformer in the distribution area. The output power of each V2G charging pile collectively meets the remaining capacity limit of the transformer, ensuring that the transformer's allowable load capacity will not be exceeded during the entire charging and discharging process of the distribution area.

[0050] After completing the objective function calculation and constraint verification, the local decision-maker outputs the final charging / discharging power command, and each V2G charging pile completes charging or discharging control according to the corresponding power. In the entire distributed decision-making mechanism, decision-making power is delegated to each local decision-maker, no longer relying on a single centralized controller to complete unified calculations. Each local decision-maker independently completes optimization calculations based on the corresponding vehicle status data and prediction results. When any local decision-maker fails, the remaining local decision-makers continue to operate independently and normally, and the overall system function is unaffected. At the same time, the failed node automatically enters a safety mode, maintaining the stable control capability of each charging node while ensuring continuous system operation.

[0051] Without altering the existing interface, a bidirectional communication link is embedded during the charging and discharging process. Vehicles and V2G charging stations exchange status data and power commands in real time via power line carrier communication (PLC) or wireless communication. The charging station's communication module supports the ISO15118-20 communication protocol stack. The ISO15118-20 standard supports bidirectional charging use cases, including V2G (Vehicle-to-Grid), V2H (Vehicle-to-Home), and V2B (Vehicle-to-Building), enabling secure and controllable bidirectional power flow between vehicles, charging stations, and the grid dispatch system. The communication process includes TLS 1.3 secure transport layer encryption, plug-and-charge authentication based on digital certificates, and negotiation of bidirectional power transmission parameters. The charging pile controls its operating mode through a bidirectional power conversion module based on the power command generated by the local decision-maker: when the power command is positive, the bidirectional power conversion module operates in charging mode, converting the AC power from the grid into DC power to charge the vehicle battery; when the power command is negative, the bidirectional power conversion module operates in discharging mode, converting the DC power from the vehicle battery into AC power to feed into the grid, and adjusting the discharge power in real time according to the grid frequency deviation to participate in primary frequency regulation, with a frequency regulation response time ≤100ms.

[0052] The specific implementation process is as follows: After the vehicle connects to the V2G charging station, a two-way communication link is established between the vehicle and the V2G charging station while maintaining the existing charging interface form. This enables continuous communication between the vehicle, the charging station, and the power grid dispatching system throughout the charging and discharging process. Once the two-way communication link is established, the vehicle and the V2G charging station continuously exchange operating data according to a predetermined communication mechanism. The exchanged data includes the vehicle's current operating status, charging and discharging status, and power commands generated by the local decision-maker. This ensures that the vehicle's operating status and charging and discharging control status are updated synchronously, while also providing real-time data support for subsequent switching of charging and discharging modes.

[0053] Data transmission between the vehicle and the V2G charging station is accomplished using either power line carrier communication (PLC) or wireless communication. When using PLC, the charging connection line simultaneously handles power transmission and data communication, enabling bidirectional transmission of status data and control commands along the charging line. When using wireless communication, the vehicle establishes a wireless connection with the V2G charging station, achieving real-time exchange of status data and power commands. Both communication methods maintain synchronization between vehicle operating status, charging station operating status, and control commands, ensuring real-time communication throughout the entire charging and discharging process.

[0054] The communication module supports the ISO15118-20 communication protocol stack. After establishing a communication connection, it completes data interaction between the vehicle, V2G charging station, and power grid dispatching system according to the ISO15118-20 communication protocol. The ISO15118-20 standard supports bidirectional charging application scenarios, including three bidirectional charging use cases: V2G, V2H, and V2B. Specifically, the V2G mode supports bidirectional energy interaction between the vehicle and the power grid, enabling the vehicle to participate in grid peak shaving, frequency regulation, and demand response; the V2H mode supports bidirectional power supply between the vehicle and a household, enabling the vehicle battery to supply power to household loads; and the V2B mode supports bidirectional power supply between the vehicle and a building, enabling the vehicle to provide power to building loads. The communication module completes data interaction for different bidirectional charging modes according to a unified communication protocol, forming a unified communication system between the vehicle, charging station, and power grid dispatching system, and achieving safe and controllable bidirectional power flow.

[0055] After establishing a communication connection, the entire communication process adheres to the ISO15118-20 communication protocol and employs a secure communication mechanism. First, TLS 1.3 secure transport layer encryption is implemented to establish an encrypted communication channel for data exchanged between the vehicle and the V2G charging station. This ensures that charging and discharging control commands, status data, and operational information are transmitted in an encrypted environment, enhancing data security during communication. Next, plug-and-charge authentication based on digital certificates is performed. After the vehicle completes the physical connection, its identity is verified using its digital certificate, eliminating the need for additional manual authentication and allowing the vehicle to enter the charging and discharging control process. Following authentication, the vehicle and the V2G charging station further negotiate bidirectional power transmission parameters, confirming the power parameters corresponding to subsequent charging or discharging modes. This ensures consistent charging and discharging control parameters between the two parties, providing a communication foundation for bidirectional power control.

[0056] After completing communication parameter negotiation, the local decision-maker generates charging and discharging power commands based on the prediction results and the vehicle's real-time operating status, and sends the power commands to the bidirectional power conversion module. The bidirectional power conversion module automatically determines the current operating mode based on the direction of the power command value. When the power command is positive, the bidirectional power conversion module enters charging mode, first rectifying the AC power input from the grid side, and then converting it into DC power that the vehicle battery can receive, continuously charging the vehicle battery. During the charging process, the communication module continues to maintain real-time interaction between vehicle status data and power commands, enabling the vehicle's current state of charge and charging status to be continuously fed back to the control system, while updating charging control information according to changes in operating status.

[0057] When the power command is negative, the bidirectional power conversion module automatically switches to discharge mode. The DC power output from the vehicle battery first enters the bidirectional power conversion module, then is converted into AC power through inversion, and finally fed into the power grid, enabling the vehicle to supply power to the grid. During the discharge process, the communication module continuously exchanges information on vehicle operating status, power grid operating status, and power control, ensuring that the discharge process remains consistent with the current power grid operating status.

[0058] After the vehicle enters discharge mode, the bidirectional power conversion module dynamically adjusts the discharge power based on real-time regional grid frequency deviation information. When the grid frequency changes, the communication module synchronously acquires the latest grid frequency deviation data and sends it to the local decision-maker. The local decision-maker updates the discharge power command based on the current grid frequency deviation, and the bidirectional power conversion module adjusts the vehicle's discharge power output to the grid in real time according to the updated power command, enabling the vehicle to participate in the primary frequency regulation process. Throughout the primary frequency regulation process, the discharge power can continuously adjust with changes in grid frequency, with the frequency regulation response time controlled within 100ms. This allows the vehicle to respond quickly to changes in grid frequency, improving the grid's operational stability during vehicle-grid interaction.

[0059] Throughout the entire two-way communication and collaborative execution process, the vehicle's operating status, the charging pile's operating status, the power grid's operating status, and the charging and discharging power commands are always kept in sync. The communication module completes two-way communication, security authentication, and parameter negotiation according to the ISO15118-20 communication protocol. The two-way power conversion module completes the switching between charging and discharging modes according to the power commands generated by the local decision-maker, and dynamically adjusts the discharging power in combination with the real-time power grid frequency deviation, so as to realize safe, continuous, and controllable two-way power interaction between the vehicle, V2G charging pile, and power grid dispatching system.

[0060] Blockchain nodes are deployed in V2G charging piles and smart converged terminals in the charging area, and V2G authorized smart contracts are deployed on the blockchain network. First, a blockchain-based digital identity is generated for each user. This digital identity includes a user identifier, vehicle identifier, wallet address, and charging / discharging permission information, establishing a unique correspondence between each user, vehicle, and transaction account participating in vehicle-to-grid interaction. This identity information serves as the foundation for subsequent charging / discharging authorization, identity verification, and revenue settlement. After the digital identity is established, each digital identity is associated with its corresponding vehicle, maintaining consistency throughout the vehicle's participation in vehicle-to-grid interaction. This ensures that the charging / discharging process, transaction process, and revenue distribution process all correspond to a unique user and vehicle.

[0061] After establishing digital identities, real-time load signals and vehicle battery operation data are uploaded to the blockchain distributed ledger. Real-time load signals include grid frequency, transformer area load rate, and dispatch command values, reflecting the current grid operating status and dispatch needs. Vehicle battery operation data includes charging / discharging start and end times, charging / discharging power, and cumulative charge, recording the vehicle's operational information throughout the entire charging / discharging process. Blockchain nodes continuously receive data from V2G charging piles and smart integrated terminals in the transformer area, writing it into the blockchain distributed ledger in chronological order. This ensures continuous recording of various operational data, providing a reliable data foundation for subsequent authorization decisions, revenue calculations, and data storage.

[0062] After real-time load signals and vehicle battery operation data are written to the blockchain distributed ledger, the V2G authorization smart contract automatically selects vehicles to participate in the discharge based on the uploaded real-time data and preset discharge vehicle screening rules. The screening rules include three conditions: the vehicle's battery state of charge (SOC) is not lower than the discharge threshold, the battery state of health (SOH) is not lower than 80%, and the user has authorized participation in V2G interaction. The smart contract reads the vehicle's current operating status information item by item. When the vehicle's current SOC meets the discharge conditions, the battery health meets the operating requirements, and the user has completed V2G authorization, the corresponding vehicle is identified as a target vehicle meeting the discharge conditions. For vehicles that do not meet the screening conditions, they maintain their original operating status and do not participate in the current discharge scheduling, ensuring that all vehicles participating in vehicle-to-grid interaction meet the requirements for safe operation and user authorization.

[0063] After the selection of discharge vehicles is completed, a tiered power allocation command is generated based on the grid load gap value and the selected discharge vehicles, combined with a tiered power allocation strategy. The grid load gap value reflects the current actual regulation demand of the grid, while the selected discharge vehicles reflect the vehicle-grid interaction resources that can participate in discharge regulation. Both pieces of information serve as the basis for power allocation. The smart contract generates corresponding tiered power allocation commands based on the regulation demand corresponding to the grid load gap value and the operating status of each discharge vehicle. Each discharge vehicle executes the corresponding discharge power control according to the tiered power allocation command, achieving a reasonable allocation of vehicle-grid interaction resources. Throughout the power allocation process, the operational data generated by the discharge vehicles is synchronously recorded to the blockchain distributed ledger, creating a complete data record of the power allocation process and the execution process.

[0064] After the discharge is completed, the V2G authorized smart contract automatically calculates the compensation amount based on the actual charging and discharging amount of the vehicle and executes the corresponding settlement transaction. The compensation amount calculation formula is as follows: Wherein, Compensation represents the amount of charge / discharge compensation; This indicates the actual amount of electricity discharged by the vehicle. Indicates the unit price of discharge compensation; This indicates the actual amount of electricity the vehicle is charged. This represents the charging unit price. The smart contract calculates the discharge compensation amount based on the actual discharge capacity of the vehicle and the discharge compensation unit price recorded in the blockchain distributed ledger. At the same time, it calculates the charging fee based on the actual charging capacity of the vehicle and the charging unit price. Based on the above calculation relationship, the final compensation amount is automatically generated, and the corresponding settlement transaction is then executed. This ensures that the calculation of charging and discharging revenue and the settlement of fees are completed automatically without the need for manual calculation and confirmation, thereby improving the efficiency of charging and discharging transaction processing.

[0065] After revenue settlement is completed, a data report is generated, and the complete charging and discharging interaction record is packaged and uploaded to the blockchain distributed ledger, forming an immutable evidence record. The charging and discharging interaction record includes the start and end time of charging and discharging, the amount charged and discharging, the settlement amount, and the participating users. All data is arranged in the order of transaction execution to form a complete record and is associated with the corresponding digital identity, so that each charging and discharging transaction forms a corresponding data evidence. After the data report is uploaded, the blockchain distributed ledger saves the complete transaction history, and all historical records are kept in a continuous storage state, providing a basis for subsequent revenue inquiries and historical traceability.

[0066] Users can view their earnings details and charging / discharging history in real time through a mobile application. The mobile application reads all transaction information of the corresponding user based on the data reports stored in the blockchain distributed ledger, displaying the start and end time of each charging / discharging session, the amount charged / discharged, the settlement amount, and the participation record. This allows users to obtain real-time earnings information for each vehicle-to-everything (V2X) interaction and view historical charging / discharging execution records, improving the transparency of the entire earnings settlement process.

[0067] Throughout the entire revenue settlement and data storage process, the V2G authorized smart contract not only completes the compensation amount calculation and settlement transaction, but also undertakes the automated management functions of charging and discharging execution, discharge calculation, and fee settlement. The smart contract continuously reads the operational data from the blockchain distributed ledger, and completes the selection of discharging vehicles, power allocation, compensation calculation, and transaction execution based on the real-time operating status. Data generated at each stage is synchronously recorded to the blockchain distributed ledger, ensuring that the data throughout the charging and discharging process remains intact and preserved. Ultimately, a trust mechanism covering the entire process of dispatch instruction generation, charging and discharging execution, and revenue settlement is constructed. This ensures that the data generated during vehicle-to-grid interaction has integrity, continuity, and immutability, while maintaining consistency between the revenue settlement process and the charging and discharging execution process, thereby improving the data credibility and transaction transparency of the entire vehicle-to-grid interaction process.

[0068] The technical solution will be described in detail below with reference to specific embodiments, but the scope of protection is not limited to the following embodiments.

[0069] Example 1 Taking a city with 10,000 electric vehicles participating in V2G dispatch as an application scenario, the entire vehicle-to-grid (V2G) interactive control system is deployed in the city's power distribution network. The power distribution cloud master station is deployed in the city's power grid dispatch center and connected to 100 smart converged terminals in various distribution areas via fiber optic networks. Each smart converged terminal in a distribution area manages approximately 100 V2G charging piles, and each V2G charging pile is connected to its corresponding access vehicle, forming a large-scale V2G interactive control network covering the entire area. Each V2G charging pile has a built-in ARM architecture edge computing chip as a local decision-maker and runs a lightweight MPC algorithm, enabling each charging pile to independently complete charging and discharging optimization calculations and control execution capabilities. The power distribution cloud master station is responsible for the unified coordination of all dispatchable resources within the area, the smart converged terminals in the distribution areas are responsible for equipment management and edge coordination within their respective distribution areas, and each V2G charging pile is responsible for vehicle-side charging and discharging execution, thus forming a hierarchical control system that combines cloud dispatch, edge coordination, and terminal execution.

[0070] During operation, the intelligent integrated terminal in each distribution area collects the real-time load rate of the transformers every 5 seconds and simultaneously acquires the operating status of the feeders. The collected results are uploaded to the distribution cloud master station, enabling the master station to continuously monitor the current load changes in each distribution area. As real-time operating data from each distribution area continuously converges, the distribution cloud master station generates load operating status information for the entire region, providing a real-time data source for subsequent load forecasting and dispatching decisions. Because the sampling period remains at 5 seconds, it can promptly reflect changes in the load of each distribution area, ensuring that the power grid operating status is always updated in real time.

[0071] After continuously receiving data uploaded from intelligent converged terminals in various distribution areas, the distribution cloud master station uses an OPTICS-EEMD-CLSTM-GWO hybrid prediction model to predict the grid load curve and electric vehicle charging and discharging potential for the next 15 minutes. Specifically, the OPTICS algorithm is first used to cluster and identify users with different vehicle usage patterns; then, the EEMD algorithm is used to decompose non-stationary load data; subsequently, the CLSTM model is used to perform time-series predictions for each component; and finally, the GWO algorithm is used to optimize the hyperparameters of the CLSTM model. After collaborative calculation by these four algorithms, the grid load prediction result and vehicle-grid interaction resource prediction result for the next 15 minutes are generated. The model prediction error is controlled at 6.7%, ensuring high accuracy and providing a reliable basis for subsequent dispatching.

[0072] Based on the prediction results, the distribution cloud master station generates regional-level V2G dispatch instructions. In this embodiment, the target discharge power is set to 500kW, and the response period is set to 14:00 to 14:30. After the dispatch instructions are generated, they are sent to all intelligent converged terminals in the distribution area via the fiber optic communication network. Each intelligent converged terminal in the distribution area receives the corresponding dispatch instructions according to its management scope and completes the distribution of the regional-level dispatch target to each V2G charging pile, enabling all charging equipment to participate in vehicle-to-grid interaction operation according to the unified dispatch target.

[0073] After receiving the dispatch command, each V2G charging station's local decision-maker begins collecting the current operating status of connected vehicles, including the vehicle's battery state of charge (SOC), battery state of health (SOH), and estimated downtime. Simultaneously, it establishes a charging and discharging optimization model based on predicted future load changes. Each local decision-maker independently calculates the optimal discharge power for its corresponding vehicle based on its operating status, without relying on other charging stations for unified calculations. This allows each vehicle to perform personalized discharge control based on its own operating conditions.

[0074] In the multi-objective optimization process, the weighting coefficients are set to α=0.4, β=0.3, and γ=0.3, where α corresponds to peak shaving and valley filling priority, β corresponds to battery protection priority, and γ corresponds to user travel guarantee priority. The economic efficiency of grid operation, battery cycle aging, and user travel demand all participate in the optimization calculation, ensuring that discharge control meets current grid peak shaving requirements while also considering vehicle battery life, and guarantees that vehicles have sufficient remaining power to meet user travel needs before leaving the station. Different optimization objectives maintain a dynamic balance within a unified model, enabling each vehicle to develop a discharge strategy corresponding to its own operating state.

[0075] After completing the optimization calculations, the local decision-maker uses the MPC algorithm to solve the multi-objective optimization model. In this embodiment, the control time domain is set to 10 minutes. Within the control time domain, the local decision-maker continuously adjusts the discharge power based on the prediction results and the vehicle's operating status, and generates the corresponding discharge power command. The discharge power command is sent to the vehicle via the ISO15118-20 communication protocol. During the communication process, the vehicle's identity is authenticated using the TLS1.3 secure transmission mechanism and digital certificate authentication, and a two-way communication connection is established, enabling the vehicle to complete the corresponding discharge control according to the discharge power command.

[0076] After the vehicle enters discharge mode, the V2G charging station controls the bidirectional power conversion module according to the discharge power command to convert the DC power output from the vehicle battery into AC power and feed it into the grid, continuously participating in grid peak shaving and frequency regulation control. Throughout the discharge process, the vehicle's operating status, the charging station's operating status, and the grid's operating status maintain real-time communication. The local decision-maker continuously adjusts the discharge power based on the real-time status, ensuring stable execution of discharge control throughout the entire response period.

[0077] Once the response period ends, each V2G charging station ceases discharge control and sends the entire charging and discharging process data to the blockchain network. The V2G authorized smart contract deployed on the blockchain network automatically calculates the compensation amount based on the vehicle's actual discharge capacity, completes the revenue settlement, and synchronously records the start and end times of charging and discharging, the cumulative discharge capacity, the settlement amount, and the information of the participating vehicles to the blockchain distributed ledger, ensuring that the entire transaction process forms an immutable data record.

[0078] Test results show that the overall system scheduling response time is less than or equal to 200ms. Even with 10,000 vehicles participating in the scheduling simultaneously, the overall system performance did not decline, maintaining stable scheduling capabilities. The daytime peak load reduction rate reached 18.7%, effectively reducing the peak load pressure on the power distribution system. The average revenue per discharge for users reached 8.5 yuan, increasing the enthusiasm of vehicles to participate in V2G interaction. The incremental cost of battery cycle aging was controlled within 0.05 yuan / kWh, indicating that the charge and discharge control takes into account both vehicle battery life and grid operation requirements, achieving efficient and coordinated scheduling of large-scale electric vehicle resources.

[0079] Example 2 The edge autonomy mode was verified using an interruption in fiber optic communication between the distribution cloud master station and the intelligent converged terminals in the distribution areas as an operational scenario. When communication between the distribution cloud master station and the intelligent converged terminals in the distribution areas is normal, all scheduling tasks are coordinated uniformly by the distribution cloud master station, and each intelligent converged terminal in the distribution areas is responsible for executing regional-level scheduling strategies. When fiber optic communication is interrupted, the intelligent converged terminals in the distribution areas automatically detect the change in communication status and immediately switch to edge autonomy mode, so that the distribution areas under their jurisdiction can continue to maintain normal charging and discharging scheduling capabilities.

[0080] After entering edge autonomy mode, the intelligent converged terminal of the distribution area calls upon locally cached historical load data and local prediction models to independently generate dispatch instructions based on the current real-time load rate of the distribution area, no longer relying on the distribution cloud master station to provide dispatch results. Since the historical load data is already stored in the local cache, it can combine the current load change trend to form a dispatch scheme that conforms to the current operating status of the distribution area, so that the charging and discharging control within the distribution area can maintain continuous operation.

[0081] After the local scheduling command is generated, the intelligent fusion terminal in the distribution area sends the corresponding scheduling information to the V2G charging piles it belongs to. Each V2G charging pile continues to use its local decision-maker to collect the vehicle's operating status, such as SOC, SOH, and estimated downtime, and completes multi-objective optimization calculations based on the local prediction results. The local decision-maker still performs optimization solutions according to the normal operation mode, generates charging and discharging power commands based on the current vehicle status, and controls the bidirectional power conversion module to complete the corresponding charging and discharging control, enabling the vehicle to continue participating in vehicle-to-grid interaction in edge autonomy mode.

[0082] Throughout the entire edge autonomy period, each V2G charging pile continuously performs charging and discharging control, while continuously recording charging and discharging process data, including vehicle operating status, charging and discharging power, cumulative charging and discharging amount, and execution time. All data is uniformly cached and saved by the intelligent fusion terminal of the distribution area, and will be uniformly synchronized after communication is restored.

[0083] Once fiber optic communication between the distribution cloud master station and the intelligent converged terminal in the distribution area is restored, the intelligent converged terminal automatically exits the edge autonomy mode and re-establishes a data communication connection with the distribution cloud master station. Subsequently, all operational data during the edge autonomy period is uploaded to the distribution cloud master station, including historical dispatch instructions, vehicle operating status, charging and discharging execution data, and distribution area operating data, enabling the distribution cloud master station to regain a complete operational record.

[0084] At the same time, blockchain nodes deployed in the blockchain network automatically supplement all ledger records during the communication interruption based on the synchronously uploaded data, so that the charging and discharging data, revenue data and transaction data generated during the edge autonomy period continue to be stored in a complete and continuous state, without affecting the integrity and credibility of the entire blockchain ledger data.

[0085] Test results show that in edge autonomy mode, all functions of the entire system remain fully operational, and the charging and discharging scheduling efficiency is maintained at over 92% of the normal operating mode, indicating that the edge autonomy mechanism can maintain high scheduling capabilities even in the event of communication anomalies. Simultaneously, after communication is restored, all operational data is successfully synchronized to the distribution cloud master station, and the blockchain ledger synchronization integrity rate reaches 100%, ensuring that all operational data is completely preserved and restored during communication interruptions, achieving continuous and stable operation of the vehicle-to-grid interactive control system even in the event of communication anomalies.

[0086] This invention employs a software-defined charging network architecture, decoupling the control plane from physical devices. This separates the scheduling decision logic for charging resources from the underlying hardware, allowing control strategies to be optimized and updated without relying on specific device structures. The distribution cloud master station is responsible for global scheduling management, the intelligent converged terminals in the distribution areas handle edge coordination, and the V2G charging piles are responsible for specific charging and discharging execution, forming a hierarchical collaborative control system. As the number of connected vehicles continues to increase, resource allocation can be completed simply by updating the scheduling strategy of the control plane, without adjusting the underlying device structure. This improves the flexibility and scalability of the entire vehicle-to-grid interaction control system, supporting a single aggregator to manage more than 100,000 electric vehicles simultaneously participating in charging and discharging scheduling, thus improving scheduling efficiency and resource management capabilities in large-scale electric vehicle access scenarios.

[0087] This invention employs a hybrid prediction model comprised of OPTICS density clustering, EEMD, CLSTM, and GWO, forming a complete prediction system covering user behavior identification, load signal processing, time-series prediction, and model optimization. OPTICS density clustering identifies user groups with different vehicle usage patterns, providing user profiles for differentiated scheduling; EEMD decomposes non-stationary grid load signals, improving the ability to process complex load data; CLSTM extracts the time-series variation characteristics of each decomposed component, taking into account both short-term load fluctuations and long-term trend predictions; and GWO further optimizes the global hyperparameters of CLSTM, improving model prediction accuracy. The multiple algorithms form a collaborative prediction mechanism, making the prediction of future grid load demand, electric vehicle charging and discharging potential, and renewable energy power generation more accurate, with prediction errors controlled within 8%, providing reliable data support for subsequent optimized scheduling.

[0088] This invention employs a distributed decision-making mechanism, delegating charging and discharging decision-making power to local decision-makers in each V2G charging pile. Each local decision-maker independently calculates charging and discharging power based on prediction results and the vehicle's real-time status, eliminating the reliance on a centralized controller to handle all scheduling tasks and effectively removing single-point bottlenecks in centralized control architectures. Each local decision-maker can perform synchronous optimization calculations, resulting in faster scheduling response and a system response speed improvement of over 50%. When any local decision-maker fails, the remaining local decision-makers continue to operate normally, with the faulty node automatically entering a safe mode without affecting the overall scheduling function. This achieves node fault isolation and local autonomous control, further enhancing the operational stability, reliability, and continuous service capability of the entire vehicle-to-grid (V2G) interactive control system.

[0089] This invention constructs a multi-objective optimization model that simultaneously considers three optimization objectives during the charging and discharging control process: grid efficiency, battery life, and user experience. Specifically, the Cgrid term comprehensively evaluates electricity purchase costs and peak-shaving and frequency regulation benefits to achieve peak shaving and valley filling; the Cbattery term comprehensively considers battery state of charge, charge / discharge rate, and battery temperature to reduce the impact of frequent charging and discharging on battery life; and the Ruser term ensures that the vehicle's actual state of charge meets users' travel needs, improving user satisfaction. These multiple optimization objectives are calculated collaboratively within a unified model, enabling the charging and discharging power to balance grid operation requirements, vehicle operating status, and user needs, achieving a coordinated and unified benefit among the grid, users, and vehicle operators, and improving the overall economic efficiency of the vehicle-grid interaction system.

[0090] This invention is compatible with the ISO 15118-20 communication standard. Without altering the existing interface, it establishes a bidirectional communication mechanism between vehicles and V2G charging stations, supporting V2G, V2H, and V2B bidirectional charging use cases. This enables secure and controllable bidirectional power flow between vehicles, charging stations, and the power grid dispatching system. The communication process employs TLS 1.3 secure transport layer encryption, combined with plug-and-charge authentication based on digital certificates and bidirectional power transmission parameter negotiation, enhancing communication security and device interconnectivity. This allows vehicles to quickly complete identity authentication and bidirectional charging / discharging control. Furthermore, it complies with the EU AFIR regulation's requirement for charging stations to support the ISO 15118-20 standard from January 1, 2027, improving system standard compatibility and international application adaptability.

[0091] This invention deploys V2G authorized smart contracts in a blockchain network to achieve fully automated management of the entire charging and discharging transaction process. The smart contract automatically completes vehicle selection, compensation calculation, and transaction settlement based on actual vehicle charging and discharging data. Simultaneously, it records dispatch instructions, the charging and discharging execution process, and the revenue settlement results to the blockchain distributed ledger, forming an immutable data record. The entire charging and discharging transaction process requires no manual intervention in settlement, improving revenue calculation efficiency and reducing manual processing costs by 70%. Furthermore, relying on the blockchain distributed ledger ensures transparent and fully traceable transaction data, constructing a comprehensive trust mechanism covering the entire process from instruction to execution to settlement, further enhancing user trust and transaction credibility.

[0092] This invention focuses on large-scale electric vehicle (EV) vehicle-to-grid (V2G) interaction scenarios, integrating several key technologies such as software-defined charging networks, hybrid prediction models, distributed decision-making, multi-objective optimization, ISO 15118-20 bidirectional communication, and blockchain smart contracts to form a complete intelligent management and control system for large-scale V2G resources. The system can dynamically adjust charging and discharging strategies based on future grid load changes and vehicle operating status, improving the flexibility of EV cluster scheduling and resource utilization efficiency, while also enhancing load forecasting accuracy, charging and discharging control precision, and overall system economy. The entire technical solution is applicable to large-scale EV access scenarios at the city level, providing reliable support for large-scale V2G resources to participate in grid peak shaving, frequency regulation, and demand response, and possesses significant commercialization value and application prospects.

[0093] This invention provides, for example Figure 2 The intelligent vehicle-to-grid (V2G) interactive control system for large-scale electric vehicles shown includes a power distribution cloud master station, intelligent fusion terminals for distribution areas, a V2G interactive protocol conversion device, V2G charging piles, and a blockchain network. The distribution cloud master station is deployed in the power grid dispatch center to receive power grid dispatching requests, perform hierarchical aggregation of dispatchable capacity, and generate regional V2G dispatching instructions. At least one smart converged terminal for the distribution area is deployed on the distribution area side corresponding to the distribution transformer and is connected to the distribution cloud master station. It is used to receive regional V2G dispatch instructions, aggregate the operation data of charging piles in the distribution area, distribute regional V2G dispatch instructions to the corresponding V2G charging piles, and execute local dispatch decisions. At least one vehicle-to-grid (V2G) communication protocol conversion device is connected between the smart fusion terminal in the distribution area and the V2G charging pile to realize protocol conversion and communication adaptation between different communication protocols. At least one V2G charging pile is connected to the vehicle-to-grid interaction protocol conversion device. The V2G charging pile includes a local decision-maker, a bidirectional power conversion module and a communication module. The local decision-maker has a built-in multi-objective optimization model, which is used to calculate the charging and discharging power command based on real-time operating data. The bidirectional power conversion module executes charging or discharging operations according to the charging and discharging power command. The communication module is used to complete the data interaction between the V2G charging pile and the electric vehicle and the intelligent integrated terminal of the distribution area. The blockchain network communicates with V2G charging piles and smart integrated terminals in the distribution area. The blockchain network is equipped with V2G authorized smart contracts to record charging and discharging data and to execute charging and discharging transaction settlements based on the charging and discharging data.

[0094] The intelligent vehicle-to-grid (V2G) interaction control method for large-scale electric vehicles provided in this embodiment of the invention is implemented through the aforementioned intelligent V2G interaction control system for large-scale electric vehicles. For details of the specific methods and processes of the intelligent V2G interaction control system for large-scale electric vehicles, please refer to the embodiments of the intelligent V2G interaction control method for large-scale electric vehicles, which will not be repeated here.

[0095] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A smart vehicle-to-grid (V2G) interactive control method for large-scale electric vehicles, characterized in that, Includes the following steps: A software-defined charging network architecture is constructed, establishing a control plane and a data plane. The control plane includes the distribution cloud master station and the intelligent converged terminal of the transformer area, while the data plane includes V2G charging piles and electric vehicles. This achieves decoupling between the control plane and physical devices, and the control plane issues charging and discharging scheduling strategies and power commands to the data plane, while the data plane feeds back operating data to the control plane. Real-time data collection of vehicle status, charging pile operation, and power grid load is used. A hybrid prediction model combining OPTICS density clustering and EEMD-CLSTM-GWO is employed to predict future power grid load demand and electric vehicle charging and discharging potential, yielding prediction results. The prediction results and real-time vehicle status data are obtained, and the local decision-maker in the V2G charging pile calculates the optimal charging and discharging power command based on a multi-objective optimization model, and generates the corresponding charging and discharging control strategy. Based on the charging and discharging power command, the bidirectional power conversion module is controlled to perform charging or discharging operations, and bidirectional data interaction between the V2G charging pile and the electric vehicle is completed based on the ISO15118-20 communication protocol. The charging and discharging process data is written into the blockchain network, and the smart contracts deployed in the blockchain network record the charging and discharging data, automatically calculate the compensation amount based on the actual charging and discharging amount, and complete the charging and discharging transaction settlement.

2. The intelligent vehicle-to-grid (V2G) interactive control method for large-scale electric vehicles according to claim 1, characterized in that, The hybrid prediction model includes the OPTICS density clustering model, the EEMD decomposition model, the CLSTM prediction model, and the GWO optimization model. The OPTICS density clustering model is used to cluster and identify user groups with different usage patterns based on user vehicle operation characteristics. The EEMD decomposition model is used to perform multi-scale decomposition of non-stationary power grid load signals. The CLSTM prediction model is used to extract the time-series variation characteristics corresponding to each decomposition component and complete the load trend prediction. The GWO optimization model is used to globally optimize the model parameters of the CLSTM prediction model.

3. The intelligent vehicle-to-grid (V2G) interactive control method for large-scale electric vehicles according to claim 1, characterized in that, The local decision-makers in V2G charging piles adopt a distributed decision-making mechanism. Each local decision-maker independently calculates the charging and discharging power commands based on the corresponding electric vehicle status data and prediction results. The distributed decision-making mechanism does not rely on the centralized controller to complete single-node decisions. When any local decision-maker malfunctions, the remaining local decision-makers continue to operate independently and maintain the charging and discharging control of the corresponding electric vehicle according to the preset safety strategy.

4. The intelligent vehicle-to-grid (V2G) interactive control method for large-scale electric vehicles according to claim 1, characterized in that, When the bidirectional power conversion module is in the discharge working state, it adjusts the discharge power output from the vehicle to the grid according to the grid frequency deviation value, so that the electric vehicle participates in the primary frequency regulation of the grid. The response time of the primary frequency regulation is less than or equal to 100ms.

5. The intelligent vehicle-to-grid (V2G) interactive control method for large-scale electric vehicles according to claim 1, characterized in that, The V2G authorized smart contract automatically calculates the compensation amount based on the actual charging and discharging volume of the electric vehicle. The compensation amount is determined based on the discharge volume, the discharge compensation unit price, the charging volume, and the charging unit price. The calculation relationship is as follows: in, Indicates the amount of discharge. This indicates the unit price for discharge compensation. Indicates the charging capacity. This indicates the unit price for charging.

6. The intelligent vehicle-to-grid (V2G) interactive control method for large-scale electric vehicles according to claim 1, characterized in that, V2G authorized smart contracts are used to automate the management of charging and discharging authorization, charging and discharging execution, discharge quantity calculation and fee settlement processes. They also store the scheduling instructions, execution data and settlement data in the blockchain distributed ledger during the charging and discharging interaction process, forming a trusted interaction mechanism based on the entire process of instruction-execution-settlement.

7. An intelligent vehicle-to-grid (V2G) interactive control system for large-scale electric vehicles, used to implement the intelligent V2G interactive control method for large-scale electric vehicles as described in any one of claims 1-6, characterized in that, This includes a power distribution cloud master station, intelligent integrated terminals for distribution areas, vehicle-to-grid (V2G) protocol conversion devices, V2G charging piles, and a blockchain network. The distribution cloud master station is deployed in the power grid dispatch center to receive power grid dispatching requests, perform hierarchical aggregation of dispatchable capacity, and generate regional V2G dispatching instructions. At least one smart converged terminal for the distribution area is deployed on the distribution area side corresponding to the distribution transformer and is connected to the distribution cloud master station. It is used to receive regional V2G dispatch instructions, aggregate the operation data of charging piles in the distribution area, distribute regional V2G dispatch instructions to the corresponding V2G charging piles, and execute local dispatch decisions. At least one vehicle-to-grid (V2G) communication protocol conversion device is connected between the smart fusion terminal in the distribution area and the V2G charging pile to realize protocol conversion and communication adaptation between different communication protocols. At least one V2G charging pile is connected to the vehicle-to-grid interaction protocol conversion device. The V2G charging pile includes a local decision-maker, a bidirectional power conversion module and a communication module. The local decision-maker has a built-in multi-objective optimization model, which is used to calculate the charging and discharging power command based on real-time operating data. The bidirectional power conversion module executes charging or discharging operations according to the charging and discharging power command. The communication module is used to complete the data interaction between the V2G charging pile and the electric vehicle and the intelligent integrated terminal of the distribution area. The blockchain network communicates with V2G charging piles and smart integrated terminals in the distribution area. The blockchain network is equipped with V2G authorized smart contracts to record charging and discharging data and to execute charging and discharging transaction settlements based on the charging and discharging data.

8. The intelligent vehicle-to-grid (V2G) interactive control system for large-scale electric vehicles according to claim 7, characterized in that, The local decision-maker incorporates a multi-objective optimization model. This model uses grid-side operating costs, battery cycle aging costs, and user satisfaction as optimization objectives. The objective function is expressed as follows: in, This represents the overall optimization target value; This represents the grid-side operating cost, used to characterize the net cost between the electricity purchase cost and the revenue from peak shaving and frequency regulation during the charging and discharging process; This represents the cost of battery cycle aging, used to characterize the lifespan loss of a vehicle battery due to charging and discharging. It represents user satisfaction and is used to characterize the extent to which the current state of battery charge of the vehicle meets the user's travel needs; , , These represent the weighting coefficients for grid operation costs, battery cycle aging costs, and user satisfaction, respectively. The weighting coefficients are configured based on grid dispatch requirements, battery type, and user operation mode.

9. The intelligent vehicle-to-grid (V2G) interactive control system for large-scale electric vehicles according to claim 7, characterized in that, The battery cycle aging cost satisfies the following relationship: in, This represents the battery aging factor, which is determined based on the battery's state of charge (SOC), charge / discharge rate (C-rate), and battery temperature. Indicates charging power or discharging power; Indicates the duration of charging and discharging; User satisfaction satisfies the following relationship: in, This indicates the user-defined threshold for the state of charge for travel protection. Indicates the vehicle's current actual state of charge; when At that time, the user satisfaction value is 1.

10. The intelligent vehicle-to-grid (V2G) interactive control system for large-scale electric vehicles according to claim 7, characterized in that, The local decision-maker uses model predictive control or dynamic programming algorithms to solve the multi-objective optimization model and generates charging and discharging power commands based on the prediction results, vehicle status data, and grid load data. When using model predictive control, the prediction time domain is 15-30 minutes and the control time domain is 5-15 minutes. The communication module supports the ISO15118-20 communication protocol stack. Based on the ISO15118-20 communication protocol, the communication module completes bidirectional communication between the V2G charging pile and the electric vehicle. It supports bidirectional charging modes such as vehicle-to-grid V2G, vehicle-to-home V2H, and vehicle-to-building V2B. The communication process includes TLS1.3 secure transmission layer encryption, plug-and-charge authentication based on digital certificates, and bidirectional power transmission parameter negotiation, enabling the V2G charging pile to switch between charging and discharging modes according to charging and discharging power commands.

Images