A community vehicle network interaction-based virtual power plant construction and power consumption scheduling method and system

Abstract

The application provides a community vehicle network interaction-based virtual power plant construction and power consumption scheduling method and system, which acquires community load, electric vehicle cluster, electricity price and communication and other multi-source data, and calculates the schedulable capacity of the electric vehicle cluster through distributed aggregation and space-time matching to form basic scheduling parameters meeting power supply reliability constraints; then, market value optimization is carried out in combination with time-of-use electricity price to generate economic scheduling strategy; further, value-added service optimization is carried out by using auxiliary service price curve to form high-value scheduling strategy; finally, reliability verification is carried out in combination with communication quality matrix to output execution scheduling configuration. The method solves the problems that distributed electric vehicle energy storage resources are difficult to be efficiently aggregated to match community load space-time demand, difficult to deeply participate in the electricity market to realize multiple economic values, and scheduling instructions are insufficient in transmission reliability in a complex communication environment.

Description

Technical Field

[0001] This invention relates to the field of smart grids, specifically to a method and system for constructing and scheduling virtual power plants and electricity consumption based on community vehicle-to-grid interaction. Background Technology

[0002] With the deepening of energy transition and the rapid increase in the number of electric vehicles (EVs), the scale of distributed energy storage resources at the community level is growing rapidly. If these resources are in a disorderly charging state, they will exacerbate the peak-valley difference in the power grid and threaten the safe and stable operation of the distribution system. Conversely, if they can be effectively aggregated and coordinated for scheduling, they can form a virtual power plant (VPP) that can participate in grid interaction, providing key support for improving grid flexibility and promoting the consumption of renewable energy. Therefore, how to transform decentralized EV energy storage resources into controllable and reliable grid regulation capabilities has become an important research direction in the field of smart grids.

[0003] Current technologies for scheduling electric vehicle (EV) clusters in communities typically employ centralized optimization control or simple time-of-use pricing strategies. Centralized control uses a central controller to collect information from all vehicles, perform calculations, and issue scheduling instructions. Pricing-based strategies primarily guide vehicle owners to charge during off-peak hours and discharge during peak hours to achieve a single economic arbitrage objective. Some research has also explored using aggregated EV resources to provide grid ancillary services.

[0004] However, these existing methods have significant limitations: First, distributed electric vehicle energy storage resources are difficult to aggregate efficiently to match the spatiotemporal demand of community loads; centralized control is computationally complex, has poor scalability, and is difficult to handle dynamic changes in vehicle status in real time; second, it is difficult to deeply participate in the electricity market to realize multiple economic values; simple electricity price response cannot achieve synergistic optimization of energy arbitrage and ancillary service revenue, resulting in insufficient economic value mining; finally, the reliability of dispatch instructions in complex communication environments is insufficient. Existing methods mostly assume idealized communication and ignore the serious impact of actual V2X network latency and packet loss on the reliability of instruction execution, making it difficult to implement optimization strategies. Summary of the Invention

[0005] In view of the above-mentioned actual situation, this application proposes a method and system for constructing a virtual power plant and scheduling electricity consumption based on community vehicle-to-grid interaction, in order to solve the problems in the existing technology that the distributed electric vehicle energy storage resources are difficult to efficiently aggregate to match the spatiotemporal demand of community load, difficult to deeply participate in the electricity market to realize multiple economic values, and insufficient reliability of dispatch instructions in complex communication environments.

[0006] A method for constructing a virtual power plant and scheduling electricity consumption based on community vehicle-to-grid interaction, the method comprising the following steps: S1, acquire data to be processed. The data to be processed includes community load time-series characteristic data, electric vehicle real-time cluster status data, power grid time-of-use price data, and V2X communication network data. The community load time-series characteristic data includes the load demand curve during peak electricity consumption and the remaining load capacity during off-peak electricity consumption. The electric vehicle real-time cluster status data includes the real-time vehicle access status matrix, the real-time battery charge status matrix, and charging and discharging efficiency parameters. The power grid time-of-use price data includes off-peak charging price signals, peak discharging price signals, and ancillary service price curves. The V2X communication network data is a real-time communication quality matrix. S2, Distributed energy storage aggregation and load spatiotemporal matching processing are performed on real-time cluster status data of electric vehicles and time-series characteristic data of community load to obtain the basic scheduling parameter matrix of virtual electric field. The distributed energy storage aggregation and load spatiotemporal matching processing is obtained by calculating the schedulable capacity distribution of electric vehicle cluster and optimizing it with community load demand in a spatiotemporal coupling to obtain a set of scheduling parameters that meet the power supply reliability constraints. S3, by combining the virtual electric field basic scheduling parameter matrix with the grid time-of-use electricity price data, market value optimization is performed to generate an economic scheduling strategy set. The market value optimization process calls the electricity consumption behavior prediction model generated based on the regression analysis of historical electricity consumption curves, generates an economic scheduling parameter set through cost-benefit analysis algorithm, and completes the feasibility verification by combining linear programming method. S4, the economic dispatch strategy set is combined with the ancillary service price curve in the power grid time-of-use price data to perform value-added service optimization processing to generate a high-value dispatch strategy set. The value-added service optimization processing is to identify high-value service opportunities through a value mining algorithm and combine it with an efficiency constraint optimization algorithm to complete the operation efficiency optimization. S5, perform communication reliability optimization processing on the high-value scheduling strategy set combined with the real-time communication quality matrix in the V2X communication network data, and generate execution scheduling configuration. The communication reliability optimization processing completes the communication feasibility analysis through the transmission reliability assessment algorithm and completes the consistency verification in combination with the distributed consensus protocol.

[0007] Furthermore, step S2 includes the following sub-steps: S201, Based on the real-time cluster status data of electric vehicles, the schedulable capacity is calculated using a distributed weighted aggregation algorithm to obtain the schedulable capacity distribution of the dynamic energy storage array. The distributed weighted aggregation algorithm for schedulable capacity calculation uses a consensus protocol to solve the maximum chargeable and dischargeable power of each node in the cluster in real time and forms a spatiotemporally distributed capacity function. S202, based on the dispatchable capacity distribution of the dynamic energy storage array and the peak-hour load demand curve and off-hour load remaining capacity in the community load time-series characteristic data, the scheduling parameters are solved by a spatiotemporal coupling optimization algorithm to obtain the virtual electric field basic scheduling parameter matrix. The spatiotemporal coupling optimization algorithm is to establish an optimization with the goal of maximizing power supply reliability and load tracking accuracy, and to solve the optimal scheduling parameter set using a constraint programming method.

[0008] Furthermore, step S3 includes the following sub-steps: S301, based on the off-peak charging price signal and peak discharging price signal in the time-of-use electricity price data of the power grid, calls the electricity consumption behavior prediction model generated by regression analysis based on historical electricity consumption curve data, and uses a cost-benefit analysis algorithm to conduct an economic evaluation of the virtual electric field basic scheduling parameter matrix, and generates an economic scheduling parameter set; S302, combining community load time-series characteristic data, uses a linear programming algorithm to verify the feasibility of the economic scheduling parameter set and generate an economic scheduling strategy set.

[0009] Furthermore, step S4 includes the following sub-steps: S401, based on the time-series characteristics of the ancillary service price curve, uses a value mining algorithm to analyze the frequency regulation service potential of the economic scheduling strategy set, identify high-value service opportunities, and generate a service optimization strategy subset; S402 combines the charging and discharging efficiency parameters in the real-time cluster status data of electric vehicles and uses an efficiency constraint optimization algorithm to optimize the operational efficiency of a subset of service optimization strategies, generating a high-value scheduling strategy set.

[0010] Furthermore, step S5 includes the following sub-steps: S501, based on the channel characteristics of the real-time communication quality matrix, performs communication feasibility analysis on the high-value scheduling strategy set through the transmission reliability assessment algorithm, and generates a communication optimization strategy set; S502, taking into account the time-varying characteristics of the real-time communication quality matrix, uses a distributed consensus protocol to verify the consistency of the communication optimization strategy set and generate the execution scheduling configuration.

[0011] Furthermore, the distributed weighted aggregation algorithm in S201 calculates the schedulable capacity by using a consensus protocol to solve for the maximum chargeable and dischargeable power of each node in the cluster in real time, and forms a spatiotemporally distributed capacity function, including local schedulable capacity calculation processing and distributed weighted aggregation processing based on the consensus protocol; the spatiotemporal coupling optimization algorithm in S202 solves for scheduling parameters by establishing an optimization with the goal of maximizing power supply reliability and load tracking accuracy, and uses a constraint programming method to solve for the optimal scheduling parameter set, including multi-objective optimization modeling processing and constraint programming solution processing.

[0012] Furthermore, the economic evaluation of the virtual electric field basic scheduling parameter matrix by using a cost-benefit analysis algorithm in S301 includes electricity consumption behavior prediction processing and cost-benefit accounting processing based on the prediction; the feasibility verification of the economic scheduling parameter set by combining community load time-series characteristic data and using a linear programming algorithm in S302 includes linear programming objective function construction processing and constraint condition construction processing.

[0013] Furthermore, in S401, the analysis of the frequency modulation service potential of the economic scheduling strategy set based on the time-series characteristics of the ancillary service price curve, using a value mining algorithm to identify high-value service opportunities, includes service revenue capacity modeling and high-value opportunity window identification. In S402, the optimization of the service optimization strategy subset based on the charging and discharging efficiency parameters in the real-time cluster status data of electric vehicles, using an efficiency constraint optimization algorithm, includes efficiency-aware revenue correction and optimal capacity allocation adjustment. In S501, the analysis of the communication feasibility of the high-value scheduling strategy set based on the channel characteristics of the real-time communication quality matrix, using a transmission reliability assessment algorithm, includes communication link reliability modeling and scheduling instruction executability analysis. In S502, the consistency verification of the communication optimization strategy set based on the time-varying characteristics of the real-time communication quality matrix, using a distributed consensus protocol, includes distributed state synchronization and consistency decision verification.

[0014] Furthermore, after S5, it also includes: S6, which performs model prediction rolling optimization processing based on the execution scheduling configuration and the real-time cluster status data of electric vehicles to generate a real-time correction configuration control strategy. The model prediction rolling optimization processing is to construct an optimization problem with the goal of minimizing scheduling deviation and solve for the optimal control instruction set.

[0015] Furthermore, this application also discloses a virtual power plant construction and power dispatching system based on community vehicle-to-grid interaction, characterized in that the system includes: The acquisition unit is used to acquire data to be processed, which includes community load time-series characteristic data, real-time electric vehicle cluster status data, grid time-of-use electricity price data, and V2X communication network data. The community load time-series characteristic data includes the load demand curve during peak electricity consumption periods and the remaining load capacity during off-peak electricity consumption periods. The real-time electric vehicle cluster status data includes the real-time vehicle access status matrix, the real-time battery charge status matrix, and charging and discharging efficiency parameters. The grid time-of-use electricity price data includes off-peak charging price signals, peak discharging price signals, and ancillary service price curves. The V2X communication network data is a real-time communication quality matrix. The aggregation and matching unit is used to perform distributed energy storage aggregation and load spatiotemporal matching processing on real-time cluster status data of electric vehicles and community load time-series characteristic data, thereby obtaining the virtual electric field basic scheduling parameter matrix. The distributed energy storage aggregation and load spatiotemporal matching processing is to calculate the schedulable capacity distribution of the electric vehicle cluster and optimize it with the community load demand in a spatiotemporal coupling to obtain a set of scheduling parameters that meet the power supply reliability constraints. The market optimization unit is used to perform market value optimization processing by combining the virtual electric field basic scheduling parameter matrix with the grid time-of-use electricity price data, and generate an economic scheduling strategy set. The market value optimization processing is to generate an economic scheduling parameter set by a cost-benefit analysis algorithm and complete the feasibility verification by combining linear programming method. The value-added optimization unit is used to perform value-added service optimization processing on the economic dispatch strategy set combined with the ancillary service price curve in the power grid time-of-use price data to generate a high-value dispatch strategy set. The value-added service optimization processing is to identify high-value service opportunities through a value mining algorithm and combine it with an efficiency constraint optimization algorithm to complete the operation efficiency optimization. The communication optimization unit is used to perform communication reliability optimization processing on the high-value scheduling strategy set combined with the real-time communication quality matrix in the V2X communication network data, and generate the execution scheduling configuration. The communication reliability optimization processing completes the communication feasibility analysis through the transmission reliability assessment algorithm and completes the consistency verification in combination with the distributed consensus protocol.

[0016] The proposed method and system for constructing and scheduling virtual power plants based on community vehicle-to-grid interaction in this application realizes the precise aggregation and spatiotemporal optimization scheduling of community distributed energy storage resources, improving power supply reliability and grid regulation capabilities; it maximizes the multiple economic benefits of virtual power plants participating in the energy market and ancillary services market; and through communication reliability assurance, it ensures the highly reliable execution of the optimized scheduling strategy. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the method flow for constructing a virtual power plant and scheduling electricity consumption based on community vehicle-to-grid interaction proposed in this application; Figure 2 A schematic diagram of a virtual power plant construction and power dispatching system based on community vehicle-to-grid interaction is provided for embodiments of this application; Detailed Implementation

[0018] The simulation technology route in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0020] The features and performance of the present invention will be further described in detail below with reference to embodiments. Please refer to the appendix. Figure 1 As shown, a method for constructing a virtual power plant and scheduling electricity consumption based on community vehicle-to-grid interaction is presented. The method includes the following steps: S1, acquire data to be processed. The data to be processed includes community load time-series characteristic data, electric vehicle real-time cluster status data, power grid time-of-use price data, and V2X communication network data. The community load time-series characteristic data includes the load demand curve during peak electricity consumption and the remaining load capacity during off-peak electricity consumption. The electric vehicle real-time cluster status data includes the real-time vehicle access status matrix, the real-time battery charge status matrix, and charging and discharging efficiency parameters. The power grid time-of-use price data includes off-peak charging price signals, peak discharging price signals, and ancillary service price curves. The V2X communication network data is a real-time communication quality matrix. In some implementations, the acquisition of data to be processed involves synchronously collecting and structurally defining multi-source heterogeneous data streams from the physical system and the electricity market through sensing, communication, and market interface units deployed within the community. These data streams collectively characterize the system's operating status, boundary constraints, and economic incentives. Specifically, the data to be processed encompasses four dimensions: community load time-series characteristics data, real-time electric vehicle cluster status data, grid time-of-use pricing data, and V2X communication network data. Each dimension provides indispensable input for subsequent optimization algorithms.

[0021] In this embodiment, the community load time-series characteristic data is used to accurately describe the periodic fluctuations and statistical patterns of the community's overall electricity demand over a time scale; this data includes the peak-hour load demand curve and the remaining load capacity during off-peak hours. The peak-hour load demand curve is a function defined in the time domain, denoted as... Where t is a time variable, this function numerically represents the continuous trajectory of the community's total active power demand over time during a predefined peak electricity consumption period. Its expression can be obtained through historical data fitting or short-term load forecasting algorithms, and it is the core basis for the virtual power plant's peak shaving and valley filling decisions. The remaining load capacity during the off-peak electricity consumption period is a scalar or time-varying function, denoted as... This represents the remaining capacity of local distributed generation (if any) that exceeds the community's load demand during off-peak hours. This capacity defines the maximum power limit that the virtual power plant can schedule for charging of electric vehicle clusters, i.e., it must meet the following constraints: ,in Total charging power for the cluster.

[0022] In some implementations, the real-time cluster status data of electric vehicles aims to dynamically sense and quantify the real-time dispatchability of distributed energy storage resources; this data includes a real-time vehicle access status matrix, a real-time battery state of charge matrix, and charge / discharge efficiency parameters. The real-time vehicle access status matrix is ​​a binary matrix, denoted as... Where M is the number of monitored time points, N is the total number of electric vehicles in the cluster, and the matrix elements are... This indicates whether the j-th electric vehicle at time t is connected to the community charging facility and is in a dispatchable state (1 indicates connected and dispatchable, 0 indicates disconnected and not dispatchable). The real-time battery state-of-charge matrix is ​​a real number matrix, denoted as . Its dimensions and Consistency, elements This represents the state of charge (SBC) of the vehicle's battery at time t, which determines the vehicle's charge / discharge potential and energy boundary. The charge / discharge efficiency parameters include charging efficiency. and discharge efficiency ,in , This parameter is used to quantify losses during energy conversion; for example, the electrical energy absorbed by an electric vehicle from the grid. With the energy stored in the battery The relationship between them is Conversely, this relationship forms the basis for accurately calculating operating costs and revenues.

[0023] Furthermore, the time-of-use electricity price data provides external price signals that drive virtual power plants to participate in market transactions and achieve economic operation; this data includes off-peak charging price signals, peak discharge price signals, and ancillary service price curves. The off-peak charging price signal is a time series function, denoted as... The value represents the electricity price charged by the power grid during officially defined off-peak hours (usually at night). The virtual power plant decision algorithm will prioritize charging during this period for lower cost. The peak discharge price signal is a time series function, denoted as . The value represents the high electricity price implemented by the power grid during peak load periods (usually in the evening). The virtual power plant decision algorithm will prioritize discharging and selling electricity during this period to generate revenue. The ancillary service price curve is a time series function, denoted as . This signal represents the price per unit of power paid by grid operators for ancillary services such as frequency regulation and reserve. This signal opens up a new source of revenue for virtual power plants beyond energy arbitrage, and its optimization objective function will incorporate a form of... The items, of which The standby capacity provided by the virtual power plant at time t.

[0024] Preferably, the V2X communication network data is the foundation for ensuring reliable and real-time transmission of control commands between the virtual power plant aggregator and the distributed electric vehicles, and its specific form is a real-time communication quality matrix; this matrix is ​​denoted as... Its elements The performance metrics characterize the communication link between the aggregator and the j-th electric vehicle at time t, including but not limited to channel bandwidth. Transmission delay Packet loss rate and signal-to-noise ratio The inherent correlation between these metrics can be described by communication models, for example, based on Shannon's theorem, the instantaneous achievable data transmission rate of the link. Can be modeled as This matrix serves as a direct input to subsequent communication reliability optimization steps, used to assess the feasibility of scheduling instruction transmission and select the optimal communication path for instruction allocation.

[0025] S2, Distributed energy storage aggregation and load spatiotemporal matching processing are performed on the real-time cluster status data of electric vehicles and the time-series characteristic data of community load to obtain the basic scheduling parameter matrix of virtual electric field. The distributed energy storage aggregation and load spatiotemporal matching processing is performed by calculating the schedulable capacity distribution of electric vehicle clusters and optimizing it with community load demand in a spatiotemporal coupling to obtain a set of scheduling parameters that meet the power supply reliability constraints. Specifically, this step includes the following sub-steps: S201, Based on the real-time cluster status data of electric vehicles, the schedulable capacity is calculated using a distributed weighted aggregation algorithm to obtain the schedulable capacity distribution of the dynamic energy storage array. The distributed weighted aggregation algorithm for schedulable capacity calculation uses a consensus protocol to solve the maximum chargeable and dischargeable power of each node in the cluster in real time and forms a spatiotemporally distributed capacity function. Specifically, the distributed weighted aggregation algorithm for schedulable capacity calculation employs a consensus protocol to solve for the maximum charge / discharge power of each node in the cluster in real time, forming a spatiotemporally distributed capacity function. This function includes local schedulable capacity calculation processing and distributed weighted aggregation processing based on the consensus protocol. The local schedulable capacity calculation processing involves each node calculating its instantaneous charge / discharge power boundary based on its own real-time state. The distributed weighted aggregation processing based on the consensus protocol involves each node exchanging information through local communication and collaboratively calculating the global schedulable capacity.

[0026] In some implementations, the local schedulable capacity calculation process forms the basis for the schedulable capacity calculation of the dynamic energy storage array; this process occurs at each distributed computing node (i.e., each electric vehicle connected to the network or its agent controller), and its input is the real-time status data of that node obtained in step S1, including its access status. With battery state of charge Specifically, at each decision time t, each node j calculates its maximum acceptable charging power based on its physical constraints. With maximum discharge power The calculations must meet the hard safety constraints of the battery management system, which are stated as follows: , ,in This indicates the upper limit of the rated power of the on-board charger for vehicle j. This indicates the upper limit of the rated discharge power of the vehicle's on-board inverter. This represents the total energy capacity of battery j in vehicle j. and These represent the safe operating limits for the battery's state of charge (e.g., 0.9 and 0.2), where Δt is the scheduling time interval. This refers to the charge / discharge efficiency parameter defined in step S1. This calculation ensures that any scheduling strategy remains within the safe operating range of individual battery cells. It should be noted that while both steps S201 and S402 involve the use of the charge / discharge efficiency parameter in the real-time cluster status data of electric vehicles, their application purposes, processing stages, and optimization goals are fundamentally different. They constitute a progressive collaborative relationship rather than being repetitive or contradictory. Step S201 focuses on the objective evaluation and aggregation of underlying physical capabilities, while step S402 focuses on the refinement and optimization of upper-level economic strategies.

[0027] In this embodiment, the distributed weighted aggregation processing based on the consensus protocol is the core of achieving accurate perception of the overall schedulable capacity of the cluster. The purpose of this processing is to enable all distributed nodes to perceive the overall schedulable charging capacity of the cluster while protecting the privacy of individual vehicles and without relying on a central coordinator. and schedulable discharge capacity A consensus is reached. It should be noted that the consensus protocol is a distributed iterative algorithm that allows each node to exchange only a small amount of data with its communication neighbors, and through multiple iterations, converges the state values ​​of all nodes to the same global value. Specifically, each node j initializes a local state variable; for example, for aggregated charging capacity, its initial value is set to... Subsequently, in each iteration k, node j determines the state value it receives from its neighbor node i. The state value is updated according to the weighted average consensus protocol. ,in This represents the set of neighboring nodes of node j. The weight coefficients assigned by node j to information from node i, and all weight coefficients satisfy the following: This ensures the convergence of the algorithm. After a sufficient number of iterations, the state values ​​of all nodes are... It will converge to the global average. To obtain a global sum rather than an average, the algorithm incorporates knowledge of the network size N or performs scaling during initialization or final output. .

[0028] Preferably, the schedulable capacity distribution of the dynamic energy storage array is the output of the above processing, which is essentially a two-dimensional spatiotemporal function. In the time dimension, this distribution is a discrete time series, representing the maximum total charging power that the entire electric vehicle cluster can provide as a virtual energy storage array at each time t within a future scheduling cycle (e.g., the next 24 hours, with an interval of Δt). With maximum total discharge power The distribution function can be expressed as two sets of time series: and Where T represents the scheduling time domain. This schedulable capacity distribution accurately quantifies the power regulation capability of the virtual power plant and is the core input parameter upon which the spatiotemporal matching optimization of community load is based in the subsequent step S202. It directly determines the extent to which the virtual power plant can respond to load demand.

[0029] S202, based on the dispatchable capacity distribution of the dynamic energy storage array and the peak-hour load demand curve and off-hour load remaining capacity in the community load time-series characteristic data, the scheduling parameters are solved by a spatiotemporal coupling optimization algorithm to obtain the virtual electric field basic scheduling parameter matrix. The spatiotemporal coupling optimization algorithm is to establish an optimization with the goal of maximizing power supply reliability and load tracking accuracy, and to solve the optimal scheduling parameter set using a constraint programming method.

[0030] Specifically, the spatiotemporal coupling optimization algorithm solves for scheduling parameters by establishing an optimization with the objectives of maximizing power supply reliability and load tracking accuracy, and uses a constraint programming method to solve for the optimal scheduling parameter set, including multi-objective optimization modeling and constraint programming solution. The multi-objective optimization modeling process transforms the dual objectives of maximizing power supply reliability and load tracking accuracy into a mathematical optimization model. The constraint programming solution process solves the optimization model under all physical and operational constraints to obtain the Pareto optimal solution set.

[0031] In some implementations, the input to the multi-objective optimization modeling process is the dynamic energy storage array schedulable capacity distribution output in step S201. and And the community load time-series characteristic data obtained in step S1, namely the load demand curve during peak electricity consumption. and the remaining capacity of the load during off-peak electricity consumption periods Specifically, the optimization model aims to simultaneously maximize power supply reliability R and load tracking accuracy P. Power supply reliability R is defined as the ability of a virtual power plant to successfully offset load demand during peak hours, and its expression is: ,in This represents the set of peak periods. Load tracking accuracy P is defined as the degree of matching between the virtual power plant charging behavior and the remaining capacity of the community during off-peak periods, and its expression is: ,in Given the set of periods of low electricity consumption, the objective function of the optimization problem can be constructed as minimizing the weighted sum of these two objectives or finding its Pareto front, i.e. ,in and The weighting coefficients reflect the different emphases on power supply reliability and smooth access.

[0032] In this embodiment, the constraint programming solution process is the process of finding feasible solutions for the aforementioned multi-objective optimization model. It should be noted that the constraints mainly originate from the cluster capacity boundary calculated in step S201 and the operational limits of the community power grid, and are expressed as follows: Power balance constraints ,in The base load at time t; cluster power capacity constraints. and and community power grid capacity constraints ,in This represents the actual charging power of the j-th electric vehicle at time t, where t is a point in time within the entire scheduling cycle. The peak-hour total load demand curve for the community is defined in step S1. N is a predefined set of peak electricity consumption periods, where N is the total number of electric vehicles in the cluster. The remaining capacity of the community power grid during the off-peak electricity consumption period as defined in step S1. The set of predefined off-peak electricity consumption periods is defined. The constrained programming method, by introducing Lagrange multipliers or employing numerical algorithms such as the interior-point method, systematically searches for the optimal values ​​of decision variables within the feasible region constrained by the aforementioned equality and inequality constraints. This optimal value represents the optimal scheduling power for each electric vehicle at each moment. .

[0033] Preferably, the virtual electric field basic scheduling parameter matrix is ​​the final output of step S202 and also the basis for subsequent economic optimization steps; this matrix is ​​denoted as... Where M is the total number of scheduling periods and N is the total number of electric vehicles. Matrix elements This represents the initial basic scheduling power command value for the j-th electric vehicle at time t. Specifically, the elements in this matrix are derived from the optimal solution set obtained through constraint programming, i.e. The matrix encapsulates the optimal power allocation scheme for each distributed energy storage unit within the virtual power plant, prioritizing only the reliability of community power supply without considering economic factors. This matrix serves as the direct input for market value optimization in subsequent step S3, ensuring that the economic strategy is optimized based on a physically feasible solution that satisfies basic power supply guarantees.

[0034] S3, by combining the virtual electric field basic scheduling parameter matrix with the grid time-of-use electricity price data, market value optimization is performed to generate an economic scheduling strategy set. The market value optimization process calls the electricity consumption behavior prediction model generated based on the regression analysis of historical electricity consumption curves, generates an economic scheduling parameter set through cost-benefit analysis algorithm, and completes the feasibility verification by combining linear programming method. Specifically, this step includes the following sub-steps: S301, based on the off-peak charging price signal and peak discharging price signal in the time-of-use electricity price data of the power grid, calls the electricity consumption behavior prediction model generated by regression analysis based on historical electricity consumption curve data, and uses a cost-benefit analysis algorithm to conduct an economic evaluation of the virtual electric field basic scheduling parameter matrix, and generates an economic scheduling parameter set; The economic evaluation of the virtual electric field basic dispatch parameter matrix through cost-benefit analysis algorithm includes electricity consumption behavior prediction processing and prediction-based cost-benefit accounting processing. The electricity consumption behavior prediction processing is based on training a regression model based on historical data to predict the community net load curve for future periods. The prediction-based cost-benefit accounting processing is a forward-looking economic evaluation of the basic dispatch strategy by integrating time-of-use electricity price signals and prediction curves.

[0035] In some implementations, the electricity consumption prediction process is a prerequisite for economic assessment. This process invokes an electricity consumption prediction model trained based on historical community load time-series characteristic data. This model is a supervised learning regression model, whose inputs are multi-dimensional feature vectors such as historical load data for the same period, date type, and weather characteristics. Its output is the community's net load demand for the future scheduling cycle. The predicted value. Specifically, this prediction model establishes a mapping function between characteristics and load values ​​by analyzing electricity consumption patterns, periodicity, and trend characteristics in historical data. ,now that ,in Let be the eigenvector at time t. This is the random error term. The predicted value... Compared to the static time-series characteristic data used in step S1, it can more accurately reflect the actual future power supply and demand scenarios, providing a forward-looking data foundation for economic assessment.

[0036] In this embodiment, the prediction-based cost-benefit analysis is the core of the evaluation; this analysis uses the predicted net load curve output by the electricity consumption behavior prediction model. Time-of-use electricity price signal and And the virtual electric field basic scheduling parameter matrix output in step S202 Both are used as input. It should be noted that the cost-benefit analysis algorithm here acts as a "simulation," and does not directly employ... Instead of performing calculations, the basic scheduling strategy is placed within the predicted load scenario. Next, evaluate its economic efficiency after execution. The algorithm traverses each scheduling instruction in the matrix. And determine its economic value based on the predicted load scenario: if the predicted load peak is at time t (i.e. (Maximum), then the discharge behavior marginal revenue It is given higher weight; if the predicted load is low, then the charging behavior... marginal cost It was given special consideration, among which The charging efficiency of vehicle j, in addition to the cost of electricity purchase, also takes into account the cost of energy loss during the charging process. For discharge efficiency, this factor considers the impact of energy loss during discharge on the saleable electricity. The algorithm's output is a set of economical scheduling parameters that take into account future market conditions and electricity demand. Each element It is itself a tuple containing physical commands and economic attributes, that is... Specifically, in this set, t represents a specific point in time, and j represents a specific electric vehicle. Each unique (t, j) combination corresponds to a tuple. It encapsulates all the key parameters for scheduling vehicle j at time t; Directly inherited from the basic scheduling parameter matrix elements in It represents the physical power command (unit: kilowatt, kW) required from vehicle j at time t to ensure basic power supply reliability in the community. A negative value indicates a charging command, and a positive value indicates a discharging command. This parameter represents the marginal cost of the instruction. Its calculation depends on the nature of the instruction: if (Charging command), then the marginal cost is expressed as This represents the additional electricity cost incurred for each additional kilowatt of charging power; if it is a discharge command, this item is usually zero or ignored. This parameter represents the marginal benefit of the instruction. Its calculation also depends on the nature of the instruction: if (Discharge command), then the marginal benefit is expressed as This represents the additional electricity sales revenue generated by each additional kilowatt of discharge power; if it is a charging instruction, this item is zero. In summary, this data structure adds its microeconomic attributes (marginal cost and marginal revenue) to each basic physical scheduling instruction, thereby upgrading the scheduling problem from a pure physical power allocation to an optimization problem that includes economic value, and providing objective function coefficients for the next step of economic optimization based on linear programming.

[0037] In an embodiment, the economic scheduling parameter set This is the output of step S301; this parameter set encapsulates the expected microeconomic characteristics corresponding to each basic dispatch instruction based on future load forecasts, under the premise of satisfying basic power supply reliability. The evaluation process integrating the forecast model ensures that the economic parameter set is no longer a static snapshot based on past experience, but a dynamic prediction of future market conditions. This provides a more forward-looking and accurate optimization objective for the linear programming algorithm in step S302, thereby generating a more economically superior set of dispatch strategies.

[0038] S302, combining community load time-series characteristic data, uses a linear programming algorithm to verify the feasibility of the economic scheduling parameter set and generate an economic scheduling strategy set; Specifically, the feasibility verification of the economic scheduling parameter set by combining community load time-series characteristic data with linear programming algorithm includes linear programming objective function construction and constraint condition construction. The linear programming objective function construction is to integrate marginal cost and marginal revenue data in the economic scheduling parameter set with the goal of maximizing net economic benefits. The constraint condition construction is to establish the feasible domain boundary of the optimization problem based on community load time-series characteristic data and physical system limits.

[0039] In some implementations, the linear programming objective function construction process is the core of step S302 in achieving economic optimization; the input to this process is the set of economic scheduling parameters output in step S301. ,in Specifically, the objective function of the linear programming algorithm aims to maximize the total net benefit throughout the entire virtual power plant scheduling cycle, and its expression is constructed as follows: It should be noted that decision variables and These represent the discharge and charging power commands for vehicle j at time t, respectively, and are specified as follows: , Coefficients in the objective function and Directly derived from the economic scheduling parameter set The elements corresponding to time and vehicle are included, which makes the objective function directly reflect the microeconomic value of each scheduling decision, guiding optimization towards the direction of maximizing net profit.

[0040] In this embodiment, the constraint construction process is crucial to ensuring that the optimization result simultaneously meets the community's electricity demand and the safe operation of the physical system; the core input of this process is the community load time-series characteristic data obtained in step S1, i.e., the load demand curve during peak electricity consumption periods. and the remaining capacity of the load during off-peak electricity consumption periods The constraints mainly include the following three categories: community power balance constraints. ,in For a predefined set of peak electricity consumption periods, this constraint ensures that during peak periods, the net discharge power of the virtual power plant can at least meet the portion of peak load exceeding base load; community capacity acceptance constraint. ,in For a predefined set of off-peak electricity consumption periods, this constraint ensures that during these periods, the total charging power of the virtual power plant does not exceed the remaining capacity limit of the community power grid; physical device power constraints. ,in and Inherited from the instantaneous power limit calculated for each vehicle j in step S201, linear programming requires setting independent constraint boundaries for each decision variable (i.e., each vehicle). These constraints ensure that the optimized instructions do not exceed the physical capability limits of each electric vehicle.

[0041] In this embodiment, the economic scheduling strategy set is the output of step S302 and also serves as the basis for subsequent participation in ancillary service market optimization; this strategy set is represented as an optimized scheduling instruction matrix. Matrix elements The value of is the optimal decision variable value obtained by solving the linear programming model. This matrix, while satisfying all community load constraints and physical equipment constraints, maximizes the net economic benefit (NEB), representing the final optimization of the economic evaluation in step S301. The economic scheduling strategy set... It will be sent to step S4 as input for value-added service optimization processing to further explore its potential to participate in the ancillary service market such as frequency regulation and standby.

[0042] S4, the economic dispatch strategy set is combined with the ancillary service price curve in the power grid time-of-use price data to perform value-added service optimization processing to generate a high-value dispatch strategy set. The value-added service optimization processing is to identify high-value service opportunities through a value mining algorithm and combine it with an efficiency constraint optimization algorithm to complete the operation efficiency optimization. Specifically, this step includes the following sub-steps: S401, based on the time-series characteristics of the ancillary service price curve, uses a value mining algorithm to analyze the frequency regulation service potential of the economic scheduling strategy set, identify high-value service opportunities, and generate a service optimization strategy subset; Specifically, the process of analyzing the frequency regulation service potential of the economic dispatch strategy set based on the time-series characteristics of the ancillary service price curve and identifying high-value service opportunities through a value mining algorithm includes service revenue capacity modeling and high-value opportunity window identification. The service revenue capacity modeling process quantifies the frequency regulation service capacity and potential revenue that a virtual power plant can provide under the economic dispatch strategy. The high-value opportunity window identification process selects the most economically valuable service periods based on the time-series matching analysis of ancillary service price signals and revenue capacity models.

[0043] In some implementations, the service revenue capacity modeling process serves as the quantitative analysis basis for value mining in step S401; the input to this process is the economic scheduling strategy set output in step S302. The schedulable discharge capacity of the dynamic energy storage array described in step S201 and the ancillary service price curve in the time-of-use electricity price data of the power grid obtained in step S1. Specifically, the value mining algorithm first calculates the remaining reserve capacity of the virtual power plant that can still provide frequency regulation services at each time t after the implementation of the economic dispatch strategy. The calculation of this capacity needs to consider the physical limits of the electric vehicle cluster and the occupancy of existing economic dispatch instructions; its expression is as follows: ,in The power values ​​of all discharge commands in the economic dispatch strategy were extracted. The value is 1 when a discharge command is issued, and 0 otherwise. This formula calculates the remaining power capacity, which can flexibly respond to grid frequency regulation commands, after subtracting the planned discharge power for energy arbitrage from the total discharge capacity. Furthermore, the algorithm calculates the potential benefit of providing this reserve capacity at time t. .

[0044] It should be noted that the ancillary service price curve used in step S401 This is fundamentally different from the electricity price signal used for energy arbitrage in step S301; This represents the price per unit of power paid by the power grid for purchasing capacity-type services or regulation services such as frequency regulation and standby. This revenue comes from providing the power grid with readily available regulation capabilities, rather than from the actual buying and selling of energy. Its revenue model is completely different from energy arbitrage and is the key for virtual power plants to tap into the high-level market value of aggregated resources.

[0045] In this embodiment, the high-value opportunity window identification process is a decision optimization process based on the aforementioned revenue model. The core of this process lies in traversing every time period within the scheduling cycle to identify those periods where ancillary service prices are high and virtual power plants have sufficient reserve capacity—that is, high-value service opportunity windows. The value mining algorithm defines a value function. The value of this function at each time t is related to the potential return. Proportional. The algorithm then sets a value threshold. This threshold can be set based on historical data or expected return targets. For all conditions that meet the threshold... The algorithm marks time periods t as high-value service opportunities and assigns these time periods and their corresponding callable capacity. Integrate them into a single set.

[0046] In this embodiment, the service optimization strategy subset is the output of step S401, and this subset is represented as a set of binary tuples. Where t represents the time point at which a service is identified as having high-value FM service potential. This represents the reserve capacity that the virtual power plant plans to commit to the ancillary services market at that moment, and The service optimization strategy subset Encapsulates additional ancillary service market opportunities that the virtual power plant can participate in, provided that the original energy arbitrage plan is basically executed. This subset will be sent to step S402 for operational efficiency optimization to generate a high-value dispatch strategy set.

[0047] S402, combining the charging and discharging efficiency parameters in the real-time cluster status data of electric vehicles, the efficiency constraint optimization algorithm is used to optimize the operational efficiency of the service optimization strategy subset, generating a high-value scheduling strategy set; Specifically, the process of optimizing the service optimization strategy subset using an efficiency constraint optimization algorithm by combining the charging and discharging efficiency parameters from the real-time cluster status data of electric vehicles includes efficiency-aware benefit correction processing and optimal capacity allocation adjustment processing. The efficiency-aware benefit correction processing calculates the energy cost required to provide ancillary services based on the charging and discharging efficiency parameters and corrects the potential benefits. The optimal capacity allocation adjustment processing prioritizes the allocation of ancillary service capacity to vehicle units with higher charging and discharging efficiency based on the corrected net benefit index.

[0048] In some implementations, the efficiency-aware benefit correction process is a computational step in step S402 that optimizes operational efficiency; the input to this process is a subset of service optimization strategies output in step S401. and the charging and discharging efficiency parameters in the real-time cluster status data of electric vehicles obtained in step S1. Specifically, the efficiency-constrained optimization algorithm identifies the additional energy loss costs that may result from providing frequency regulation services; when the virtual power plant commits to providing... When providing frequency regulation services for capacity, to cope with potential downward adjustments from the grid (increasing charging or decreasing discharging), a corresponding energy buffer needs to be reserved. The charging cost of this buffer energy increases due to efficiency losses. Similarly, the energy reserved to cope with upward adjustments (increasing discharging or decreasing charging) will also have its actual discharge capacity reduced due to efficiency losses during discharge. The algorithm defines the net revenue per unit power of ancillary services provided. for: ,in The energy call rate coefficient for frequency modulation services is estimated based on historical data (dimensionless). The average charging and discharging efficiency of the cluster. The ancillary service price at time t (unit: yuan / kilowatt) is the gross revenue of this service. The prices are for off-peak and peak electricity rates (unit: yuan / kWh). This calculation represents the expected unit energy cost in response to the downward adjustment order. This item calculates the expected unit energy opportunity cost of responding to the upward adjustment order; this net benefit indicator... It is a core economic indicator for judging whether an ancillary service opportunity truly has high value. This formula shows that the price of ancillary services... The energy cost must be high enough to cover the expected energy costs incurred due to efficiency losses for the service to have real economic value.

[0049] In this embodiment, the optimal capacity allocation adjustment process aims to maximize the overall net benefit of ancillary service activities. This process refines a subset of service optimization strategies based on the results of the benefit correction process. It should be noted that this process considers not only whether to provide capacity, but also which vehicles will provide that capacity. The efficiency-constrained optimization algorithm constructs an optimization problem with the objective of maximizing overall net benefit, where the decision variable is the ancillary service capacity committed by each vehicle j at time t. The objective function is: ,in It considers the net benefit of the efficiency of vehicle j, that is The constraints include: total cluster capacity constraint. To ensure that the total commitment remains unchanged, of which This represents summing over all electric vehicle indices j in the cluster. This represents the frequency regulation service commitment power to be allocated to the j-th electric vehicle at time t. This represents the total frequency regulation service power that the virtual power plant plans to commit to the ancillary services market at time t; individual capacity constraints. Ensure that the remaining dispatchable capacity of a single vehicle is not exceeded. This represents the maximum remaining discharge power of the j-th electric vehicle that can be used to provide frequency regulation services after executing the economic dispatch instruction at time t. Specifically, it is expressed as... ,in Inherited from step S201, it represents the physical maximum discharge power of vehicle j at time t. The economic dispatch strategy set inherited from step S302 represents the planned energy arbitrage discharge instructions. This function is used to extract discharge commands (positive values), and zero if the command is to charge (negative values). This constraint ensures that the ancillary service capacity allocated to vehicle j will never exceed its current actual physical capacity limit. This optimization problem essentially treats ancillary service capacity as a resource, prioritizing its allocation to vehicles with higher charging and discharging efficiency, thus providing a greater net benefit from the service.

[0050] In this embodiment, the high-value scheduling strategy set is the output of step S402 and also represents the final economic solution of the entire power dispatching method; this strategy set is characterized as an extended scheduling instruction matrix. The first N columns of this matrix are inherited from the economic scheduling strategy set. ,element For j=1,2,...,N, this represents the basic energy arbitrage instruction. The newly added (N+1)th column is a vector. Its elements represents the total capacity of auxiliary services committed to the entire cluster at every time t. The high-value scheduling strategy set... Simultaneously, it encapsulates the decision to maximize energy arbitrage profits and ancillary service profits, representing a high degree of unity between physical feasibility and economic optimality. This matrix will be sent to step S5 for communication reliability optimization to ensure its reliable execution.

[0051] S5, perform communication reliability optimization processing on the high-value scheduling strategy set combined with the real-time communication quality matrix in the V2X communication network data to generate the execution scheduling configuration. The communication reliability optimization processing is to complete the communication feasibility analysis by the transmission reliability assessment algorithm and complete the consistency verification by the distributed consensus protocol. Specifically, this step includes the following sub-steps: S501, based on the channel characteristics of the real-time communication quality matrix, performs communication feasibility analysis on the high-value scheduling strategy set through the transmission reliability assessment algorithm, and generates a communication optimization strategy set; Specifically, the channel characteristics based on the real-time communication quality matrix, and the communication feasibility analysis of the high-value scheduling strategy set through the transmission reliability assessment algorithm, include communication link reliability modeling and scheduling instruction executability analysis. The communication link reliability modeling quantifies the real-time data transmission reliability of each link in the V2X communication network. The scheduling instruction executability analysis evaluates the transmission success probability of the scheduling instruction set based on the link reliability model and identifies communication bottlenecks.

[0052] In some implementations, the communication link reliability modeling process is the basis for the communication feasibility analysis in step S501; the input to this process is the real-time communication quality matrix in the V2X communication network data obtained in step S1. Specifically, the transmission reliability assessment algorithm extracts key channel characteristic parameters, including channel bandwidth, for the communication link between the aggregator and each electric vehicle j. Transmission delay Packet loss rate and signal-to-noise ratio The algorithm establishes a real-time data transmission reliability model for each link based on these parameters. This model is a comprehensive indicator, and its expression can be constructed as follows: ,in and These are weighting coefficients determined based on communication protocols and service requirements. The item characterizes the reliability of the link. The item characterizes latency sensitivity (the greater the latency, the lower the reliability score). The term characterizes channel quality (the higher the signal-to-noise ratio, the higher the reliability score). The model outputs a scalar between 0 and 1, quantitatively describing the probability of successfully transmitting data through the link at time t.

[0053] In this embodiment, the scheduling instruction executability analysis process is a key step in combining the communication reliability model with specific scheduling services; the input to this process is the aforementioned link reliability model. and the high-value scheduling strategy set output by step S402 It should be noted that the transmission reliability assessment algorithm needs to calculate the successful transmission probability of the entire scheduling instruction set; since a scheduling instruction typically contains multiple data packets, and its successful transmission requires all data packets to be correctly received, the probability of successful execution of an instruction sent to vehicle j is calculated. It can be modeled as: Where K is the number of data packets contained in a complete scheduling instruction, and this value depends on the communication protocol and the amount of instruction information. Furthermore, the algorithm sets a reliability threshold. (e.g., 0.95); for all satisfying The algorithm marks scheduling instructions (i.e., instructions that need to be sent to vehicle j at time t) as instructions with communication risks and identifies their corresponding communication links as bottleneck links. These instructions may not be able to be executed due to unreliable communication and require optimization.

[0054] Preferably, the communication optimization strategy set is the output of step S501; this set is represented as a set of triples. Here, t and j identify scheduling instructions that pose a communication risk. This is a set of optimization actions for this instruction. These optimization actions include, but are not limited to: switching the instruction transmission path (selecting a relay node with higher reliability), and adjusting the instruction encoding and modulation scheme (adopting a more robust encoding method to improve performance). ), instruction retransmission strategy configuration (pre-set retransmission mechanism for important instructions). The communication optimization strategy set. The set of communication layer optimization measures that need to be taken to ensure the reliable execution of high-value scheduling strategies is encapsulated and will be sent to step S502, where a consensus verification will be performed through a distributed consensus protocol to generate an executable scheduling configuration.

[0055] S502, taking into account the time-varying characteristics of the real-time communication quality matrix, uses a distributed consensus protocol to verify the consistency of the communication optimization strategy set and generate the execution scheduling configuration.

[0056] Specifically, the consensus verification of the communication optimization strategy set through a distributed consensus protocol, which combines the time-varying characteristics of the real-time communication quality matrix, includes distributed state synchronization processing and consensus decision verification processing. The distributed state synchronization processing involves each node in the network exchanging local network state information based on the consensus protocol to achieve synchronization of its understanding of the global communication environment. The consensus decision verification processing involves distributed voting on the feasibility of the communication optimization strategy based on the synchronized global network state and reaching a consensus.

[0057] In some implementations, the distributed state synchronization process is the basis for the consistency verification in step S502; the input to this process is the real-time communication quality matrix obtained in step S1. The data exhibits time-varying characteristics. Specifically, the distributed consensus protocol requires the aggregator and each electric vehicle node to participate in the verification process as peer nodes; each node j maintains a local state vector. This vector contains communication quality information that it perceives itself, such as... The protocol achieves state synchronization through periodic information exchange between nodes; in each iteration k, node j synchronizes its state based on the state vector received from its neighbor node i. The update rules for the self-perception of its state follow a weighted average consensus algorithm: ,in This represents the set of neighboring nodes of node j. This represents the updated state perception of node j at iteration step k+1. The weight represents the information about node j's current state when updating its own state. The level of trust or retention rate, This represents the current state perception of node j at iteration step k. The weight represents the weight of node j when updating its own state, based on the state information received from neighbor node i. The weight or level of trust assigned. This represents the state perception of a neighboring node i of node j at iteration step k. After multiple iterations, the state vectors of all nodes are... It will converge to a consensus value. This value represents the global average communication state jointly confirmed by the network, providing a common cognitive basis for subsequent consistent decisions.

[0058] In some implementations, the consistency decision verification process ensures the executability of the communication optimization strategy at the network layer; the input to this process is the global communication state obtained through the aforementioned synchronization. and the communication optimization strategy set output in step S501 It should be noted that the verification process is for... Each optimization action (e.g., path switching, encoding adjustment) Conduct distributed feasibility assessment; each node j is based on the global state after consensus. Calculate independently the expected communication reliability gain after performing this optimization action. The protocol then initiates a round of distributed voting; each node j selects optimization actions. The condition for voting in favor is that its local calculations show... (in (For the preset gain threshold). An optimization action. A node is marked as final executable if and only if more than a preset percentage (e.g., 80%) of the participating nodes vote in favor, thereby achieving consensus in the distributed system.

[0059] Preferably, the execution scheduling configuration is the final output of step S502 and also the final executable scheme of the entire power consumption scheduling method; this configuration is represented as a set of four tuples. Where t and j identify the target and time of the scheduling instruction. Inherited from high-value scheduling strategy set This represents the final physical power command value (including energy arbitrage and ancillary service capacity). This represents the communication optimization actions determined after consistency verification to ensure reliable transmission of the instruction (empty if no optimization is needed). The execution scheduling configuration... It is the end point of the technical solution's logical chain. It integrates economic dispatch strategies that have undergone multiple rounds of optimization and communication assurance strategies tailored for it, ensuring the reliable implementation and execution of virtual power plant dispatch commands in complex network environments.

[0060] S6. Based on the execution scheduling configuration and the real-time cluster status data of electric vehicles, model prediction rolling optimization processing is performed to generate a real-time correction configuration control strategy. The model prediction rolling optimization processing is to construct an optimization problem with the goal of minimizing scheduling deviation and solve for the optimal control instruction set.

[0061] Step S6 involves model prediction rolling optimization based on the execution scheduling configuration and real-time cluster status data of electric vehicles to generate a real-time correction configuration control strategy. This step serves as a closed-loop feedback and real-time correction link in the entire optimization scheduling process, ensuring that the virtual power plant can continuously maintain optimal operation in a dynamically changing environment. The core of model prediction rolling optimization lies in addressing the uncertainty of the system state through periodic re-optimization. This is achieved by constructing an optimization problem with the goal of minimizing scheduling deviation and solving for the optimal control instruction set.

[0062] In some implementations, the model prediction rolling optimization process involves establishing a predictive model that uses the latest real-time cluster status data of electric vehicles to predict system behavior over a finite future time domain. This prediction includes changes in the access and exit states of electric vehicles, the evolution of battery state of charge, and fluctuations in community load. It should be noted that this predictive model is executed rollingly based on the actual state at the current moment, thereby incorporating the latest system information and reducing uncertainties caused by predictive model mismatch.

[0063] In this embodiment, the subsequent processing constructs an optimization problem with the objective of minimizing scheduling deviation. This scheduling deviation refers to the difference between the actual execution and the execution scheduling configuration generated in step S502, specifically including deviations in power command execution, auxiliary service capacity provision, and the resulting economic deviations. The decision variables of the optimization problem are the adjustments to various scheduling commands over the current and several future scheduling periods. By solving this optimization problem, an optimal set of control commands can be obtained. This set of commands can, under the premise of satisfying all physical and safety constraints, maximally correct the deviations caused by the discrepancy between the real-time state and the predicted state, allowing the system operating state to approximate the initially planned optimal trajectory again. Preferably, the model prediction rolling optimization processing adopts a rolling time-domain strategy; that is, at the beginning of each scheduling cycle, the prediction and optimization process is re-executed based on the latest system state, but only the commands corresponding to the current moment in the optimization result are implemented. In the next cycle, the commands for subsequent periods are ignored, and optimization restarts. This strategy continuously utilizes the latest feedback information to promptly correct decisions, forming a closed-loop control circuit of "prediction-optimization-execution-feedback." This effectively overcomes the inaccuracies of long-term predictions in the system and significantly improves the robustness and adaptability of the virtual power plant dispatching system in the face of real-time disturbances. The real-time correction configuration control strategy is the output of this rolling optimization process. It serves as a dynamic supplement and calibration to the execution dispatch configuration, and is thus distributed to each execution unit to ensure the continuous, stable, and economical operation of the virtual power plant.

[0064] Based on the description of the above embodiments of the virtual power plant construction and power dispatching method based on community vehicle-to-grid interaction, this application also discloses a virtual power plant construction and power dispatching system based on community vehicle-to-grid interaction. This system can be a computer program (including program code) that runs the aforementioned virtual power plant construction and power dispatching method based on community vehicle-to-grid interaction. Please see the appendix. Figure 2 As shown, the virtual power plant construction and power dispatching system based on community vehicle-to-grid interaction can operate the following units: The acquisition unit 110 is used to acquire data to be processed. The data to be processed includes community load time-series characteristic data, electric vehicle real-time cluster status data, power grid time-of-use price data, and V2X communication network data. The community load time-series characteristic data includes the load demand curve during peak electricity consumption and the remaining load capacity during off-peak electricity consumption. The electric vehicle real-time cluster status data includes the real-time vehicle access status matrix, the real-time battery charge status matrix, and charging and discharging efficiency parameters. The power grid time-of-use price data includes off-peak charging price signals, peak discharging price signals, and ancillary service price curves. The V2X communication network data is a real-time communication quality matrix. The aggregation and matching unit 120 is used to perform distributed energy storage aggregation and load spatiotemporal matching processing on real-time cluster status data of electric vehicles and community load time-series characteristic data, thereby obtaining a virtual electric field basic scheduling parameter matrix. The distributed energy storage aggregation and load spatiotemporal matching processing is to calculate the schedulable capacity distribution of the electric vehicle cluster and optimize it with the community load demand in a spatiotemporal coupling to obtain a set of scheduling parameters that meet the power supply reliability constraints. The market optimization unit 130 is used to perform market value optimization processing by combining the virtual electric field basic scheduling parameter matrix with the grid time-of-use electricity price data to generate an economic scheduling strategy set. The market value optimization processing is to generate an economic scheduling parameter set by a cost-benefit analysis algorithm and complete the feasibility verification by combining it with a linear programming method. The value-added optimization unit 140 is used to perform value-added service optimization processing on the economic dispatch strategy set combined with the auxiliary service price curve in the power grid time-of-use price data to generate a high-value dispatch strategy set. The value-added service optimization processing is to identify high-value service opportunities through a value mining algorithm and combine it with an efficiency constraint optimization algorithm to complete the operation efficiency optimization. The communication optimization unit 150 is used to perform communication reliability optimization processing on the high-value scheduling strategy set combined with the real-time communication quality matrix in the V2X communication network data, and generate the execution scheduling configuration. The communication reliability optimization processing completes the communication feasibility analysis through the transmission reliability assessment algorithm and completes the consistency verification in combination with the distributed consensus protocol.

[0065] The above description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be altered within the scope of the concept described herein through the above teachings or related technologies or knowledge. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims.

Claims

1. A method for constructing a virtual power plant and scheduling electricity consumption based on community vehicle-to-grid interaction, characterized in that, The method includes the following steps: S1, acquire data to be processed. The data to be processed includes community load time-series characteristic data, electric vehicle real-time cluster status data, power grid time-of-use price data, and V2X communication network data. The community load time-series characteristic data includes the load demand curve during peak electricity consumption and the remaining load capacity during off-peak electricity consumption. The electric vehicle real-time cluster status data includes the real-time vehicle access status matrix, the real-time battery charge status matrix, and charging and discharging efficiency parameters. The power grid time-of-use price data includes off-peak charging price signals, peak discharging price signals, and ancillary service price curves. The V2X communication network data is a real-time communication quality matrix. S2, Distributed energy storage aggregation and load spatiotemporal matching processing are performed on real-time cluster status data of electric vehicles and time-series characteristic data of community load to obtain the basic scheduling parameter matrix of virtual power plant. The distributed energy storage aggregation and load spatiotemporal matching processing is performed by calculating the schedulable capacity distribution of electric vehicle clusters and optimizing it with community load demand in a spatiotemporal coupling to obtain a set of scheduling parameters that meet the power supply reliability constraints. Step S2 includes the following sub-steps: S201, Based on the real-time cluster status data of electric vehicles, the schedulable capacity is calculated using a distributed weighted aggregation algorithm to obtain the schedulable capacity distribution of the dynamic energy storage array. The distributed weighted aggregation algorithm for schedulable capacity calculation uses a consensus protocol to solve the maximum chargeable and dischargeable power of each node in the cluster in real time and forms a spatiotemporally distributed capacity function. S202, based on the dispatchable capacity distribution of the dynamic energy storage array and the peak-hour load demand curve and off-hour load remaining capacity in the community load time-series characteristic data, the scheduling parameters are solved through a spatiotemporal coupling optimization algorithm to obtain the basic scheduling parameter matrix of the virtual power plant. The spatiotemporal coupling optimization algorithm is to establish an optimization with the goal of maximizing power supply reliability and load tracking accuracy, and to solve the optimal scheduling parameter set using a constraint programming method. S3. Market value optimization is performed by combining the virtual power plant basic dispatch parameter matrix with the grid time-of-use price data to generate an economic dispatch strategy set. The market value optimization process calls the electricity consumption behavior prediction model generated based on the regression analysis of historical electricity consumption curves, generates an economic dispatch parameter set through cost-benefit analysis algorithm, and completes the feasibility verification by combining linear programming method. S4, the economic dispatch strategy set is combined with the ancillary service price curve in the power grid time-of-use price data to perform value-added service optimization processing to generate a high-value dispatch strategy set. The value-added service optimization processing is to identify high-value service opportunities through a value mining algorithm and combine it with an efficiency constraint optimization algorithm to complete the operation efficiency optimization. Step S4 includes the following sub-steps: S401, based on the time-series characteristics of the ancillary service price curve, uses a value mining algorithm to analyze the frequency regulation service potential of the economic scheduling strategy set, identify high-value service opportunities, and generate a service optimization strategy subset; S402, combining the charging and discharging efficiency parameters in the real-time cluster status data of electric vehicles, optimizes the operational efficiency of the service optimization strategy subset through the efficiency constraint optimization algorithm to generate a high-value scheduling strategy set; S5, perform communication reliability optimization processing on the high-value scheduling strategy set combined with the real-time communication quality matrix in the V2X communication network data to generate the execution scheduling configuration. The communication reliability optimization processing is to complete the communication feasibility analysis by the transmission reliability assessment algorithm and complete the consistency verification by the distributed consensus protocol. Step S5 includes the following sub-steps: S501, based on the channel characteristics of the real-time communication quality matrix, performs communication feasibility analysis on the high-value scheduling strategy set through the transmission reliability assessment algorithm, and generates a communication optimization strategy set; S502, taking into account the time-varying characteristics of the real-time communication quality matrix, uses a distributed consensus protocol to verify the consistency of the communication optimization strategy set and generate the execution scheduling configuration.

2. The method for constructing a virtual power plant and scheduling electricity consumption based on community vehicle-to-grid interaction according to claim 1, characterized in that, Step S3 includes the following sub-steps: S301, based on the off-peak charging price signal and peak discharging price signal in the time-of-use electricity price data of the power grid, calls the electricity consumption behavior prediction model generated by regression analysis based on historical electricity consumption curve data, and uses the cost-benefit analysis algorithm to conduct an economic evaluation of the virtual power plant basic dispatch parameter matrix, and generates an economic dispatch parameter set; S302, combining community load time-series characteristic data, uses a linear programming algorithm to verify the feasibility of the economic scheduling parameter set and generate an economic scheduling strategy set.

3. The method for constructing a virtual power plant and scheduling electricity consumption based on community vehicle-to-grid interaction according to claim 1, characterized in that, The distributed weighted aggregation algorithm in S201 calculates schedulable capacity by using a consensus protocol to solve the maximum chargeable and dischargeable power of each node in the cluster in real time, and forms a spatiotemporally distributed capacity function, including local schedulable capacity calculation processing and distributed weighted aggregation processing based on the consensus protocol. The time-space coupling optimization algorithm described in S202 solves the scheduling parameters by establishing an optimization with the goal of maximizing power supply reliability and load tracking accuracy, and uses a constraint programming method to solve for the optimal scheduling parameter set, including multi-objective optimization modeling and constraint programming solution processing.

4. The method for constructing a virtual power plant and scheduling electricity consumption based on community vehicle-to-grid interaction according to claim 2, characterized in that, The economic evaluation of the virtual power plant's basic dispatch parameter matrix using a cost-benefit analysis algorithm in S301 includes electricity consumption behavior prediction and cost-benefit accounting based on the prediction; the feasibility verification of the economic dispatch parameter set using a linear programming algorithm in S302, combined with community load time-series characteristic data, includes linear programming objective function construction and constraint condition construction.

5. The method for constructing a virtual power plant and scheduling electricity consumption based on community vehicle-to-grid interaction according to claim 1, characterized in that, The time-series characteristics of the ancillary service price curve described in S401, which use a value mining algorithm to analyze the frequency regulation service potential of the economic scheduling strategy set and identify high-value service opportunities, include service revenue capacity modeling and high-value opportunity window identification. The S402 step of combining the charging and discharging efficiency parameters in the real-time cluster status data of electric vehicles with the efficiency constraint optimization algorithm to optimize the service optimization strategy subset includes efficiency-aware benefit correction processing and optimal capacity allocation adjustment processing. The S501 method, based on the channel characteristics of the real-time communication quality matrix, performs communication feasibility analysis on the high-value scheduling strategy set through a transmission reliability assessment algorithm, including communication link reliability modeling and scheduling instruction executability analysis. The S502 method, combining the time-varying characteristics of the real-time communication quality matrix, performs consistency verification on the communication optimization strategy set through a distributed consensus protocol, including distributed state synchronization and consistency decision verification.

6. The method for constructing a virtual power plant and scheduling electricity consumption based on community vehicle-to-grid interaction according to claim 1, characterized in that, Step S5 is followed by S6, which involves performing model prediction rolling optimization based on the execution scheduling configuration and real-time cluster status data of electric vehicles to generate a real-time correction configuration control strategy. The model prediction rolling optimization process is achieved by constructing an optimization problem with the goal of minimizing scheduling deviation and solving for the optimal control instruction set.

7. A virtual power plant construction and power dispatching system based on community vehicle-to-grid interaction, characterized in that, The system includes: The acquisition unit is used to acquire data to be processed, which includes community load time-series characteristic data, real-time electric vehicle cluster status data, grid time-of-use electricity price data, and V2X communication network data. The community load time-series characteristic data includes the load demand curve during peak electricity consumption periods and the remaining load capacity during off-peak electricity consumption periods. The real-time electric vehicle cluster status data includes the real-time vehicle access status matrix, the real-time battery charge status matrix, and charging and discharging efficiency parameters. The grid time-of-use electricity price data includes off-peak charging price signals, peak discharging price signals, and ancillary service price curves. The V2X communication network data is a real-time communication quality matrix. The aggregation and matching unit is used to perform distributed energy storage aggregation and load spatiotemporal matching processing on real-time electric vehicle cluster status data and community load time-series characteristic data to obtain a virtual power plant basic scheduling parameter matrix. The distributed energy storage aggregation and load spatiotemporal matching processing involves calculating the dispatchable capacity distribution of the electric vehicle cluster and optimizing it in a spatiotemporal coupling with community load demand to obtain a set of scheduling parameters that meet power supply reliability constraints, including: S201, Based on the real-time cluster status data of electric vehicles, the schedulable capacity is calculated using a distributed weighted aggregation algorithm to obtain the schedulable capacity distribution of the dynamic energy storage array. The distributed weighted aggregation algorithm for schedulable capacity calculation uses a consensus protocol to solve the maximum chargeable and dischargeable power of each node in the cluster in real time and forms a spatiotemporally distributed capacity function. S202, based on the dispatchable capacity distribution of the dynamic energy storage array and the peak-hour load demand curve and off-hour load remaining capacity in the community load time-series characteristic data, the scheduling parameters are solved through a spatiotemporal coupling optimization algorithm to obtain the basic scheduling parameter matrix of the virtual power plant. The spatiotemporal coupling optimization algorithm is to establish an optimization with the goal of maximizing power supply reliability and load tracking accuracy, and to solve the optimal scheduling parameter set using a constraint programming method. The market optimization unit is used to perform market value optimization processing by combining the virtual power plant basic dispatch parameter matrix with the grid time-of-use electricity price data, and generate an economic dispatch strategy set. The market value optimization processing calls the electricity consumption behavior prediction model generated based on the regression analysis of historical electricity consumption curves, generates an economic dispatch parameter set through cost-benefit analysis algorithm, and completes the feasibility verification by combining linear programming method. The value-added optimization unit is used to perform value-added service optimization processing on the economic dispatch strategy set combined with the ancillary service price curve in the power grid time-of-use price data, generating a high-value dispatch strategy set. The value-added service optimization processing involves identifying high-value service opportunities through a value mining algorithm and combining it with an efficiency constraint optimization algorithm to optimize operational efficiency, including: S401, based on the time-series characteristics of the ancillary service price curve, uses a value mining algorithm to analyze the frequency regulation service potential of the economic scheduling strategy set, identify high-value service opportunities, and generate a service optimization strategy subset; S402, combining the charging and discharging efficiency parameters in the real-time cluster status data of electric vehicles, optimizes the operational efficiency of the service optimization strategy subset through the efficiency constraint optimization algorithm to generate a high-value scheduling strategy set; The communication optimization unit is used to perform communication reliability optimization processing on the high-value scheduling strategy set combined with the real-time communication quality matrix in V2X communication network data, and generate execution scheduling configuration. The communication reliability optimization processing completes communication feasibility analysis through a transmission reliability assessment algorithm and completes consistency verification in conjunction with a distributed consensus protocol, including: S501, based on the channel characteristics of the real-time communication quality matrix, performs communication feasibility analysis on the high-value scheduling strategy set through the transmission reliability assessment algorithm, and generates a communication optimization strategy set; S502, taking into account the time-varying characteristics of the real-time communication quality matrix, uses a distributed consensus protocol to verify the consistency of the communication optimization strategy set and generate the execution scheduling configuration.

Images