Dynamic operation optimization method and system for source-network-load-storage integrated overcharge system

Abstract

The invention discloses a dynamic operation optimization method and system for a source-grid-load-storage integrated overcharge system, and relates to the technical field of intelligent power grids. The method comprises the following steps: periodically acquiring multi-source real-time operation data of a system, and performing synchronous verification processing to form a standardized system state data set; generating an ultra-short-term prediction curve of photovoltaic power generation and charging load based on the data set and weather forecast data, and constructing a mixed integer linear programming model containing multiple constraints based on the ultra-short-term prediction curve by taking minimization of total operation cost as a target; solving the model to obtain an optimal power distribution plan of each unit in a future time period, and immediately issuing a real-time reference power instruction at the current moment to each execution device; and after a preset optimization period, a new round of complete process from data acquisition to instruction issuing is automatically started, and rolling optimization is performed by using the updated system state data, so that closed-loop dynamic optimization control is realized, and the system economy and the power grid interaction capability are improved.

Description

Technical Field

[0001] This invention relates to the field of smart grid technology, specifically to a dynamic operation optimization method and system for an integrated supercharging system with power generation, grid, load, and energy storage. Background Technology

[0002] With the deepening of the global energy transition strategy and the establishment of the "dual carbon" target, the electric vehicle industry is experiencing explosive growth. As a key infrastructure for the widespread adoption of electric vehicles, the demand for high-power supercharging stations is increasingly urgent. However, supercharging station loads are characterized by strong randomness, high volatility, and high power levels. Their large-scale integration poses a severe challenge to the stable operation of regional power distribution networks, mainly manifested in increased peak loads, voltage fluctuations, and increased line losses. To enhance the grid's capacity to absorb renewable energy and alleviate expansion pressure, combining supercharging stations with distributed photovoltaic and energy storage systems to construct an integrated source-grid-load-storage system has become an important development direction. Against this backdrop, how to coordinate and optimize the management of distributed power sources, energy storage units, and flexible charging loads within the system is the core key to achieving economical, efficient, and grid-friendly operation.

[0003] Existing technologies have conducted various studies on the operational optimization of supercharging stations. Some solutions focus on guiding users to charge in an orderly manner through electricity pricing strategies, thereby smoothing the load curve; others utilize energy storage systems for peak shaving and valley filling. However, existing technical solutions often face challenges, such as the volatility of photovoltaic output, the uncertainty of electric vehicle charging demand, and the coordinated scheduling of diversified controllable resources such as energy storage and V2G. These factors often limit the overall optimization effect and real-time control capabilities of the system. How to dynamically respond to grid demand while balancing the economic efficiency of system operation and the satisfaction of user charging needs has become an urgent problem to be solved in the optimized management of supercharging stations. Therefore, this field needs a dynamic operation optimization method that can deeply integrate the elements of power generation, grid, load, and storage, and can perform online rolling optimization to comprehensively improve the overall operational efficiency of supercharging stations and their support capabilities for the grid. Summary of the Invention

[0004] Based on the shortcomings of the prior art described above, the purpose of this invention is to provide a dynamic operation optimization method and system for an integrated supercharging system with power generation, grid, load and storage, in order to solve the above-mentioned technical problems.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a dynamic operation optimization method for an integrated supercharging system comprising:

[0006] S1: Periodically acquire multi-source real-time operating data of the supercharging system, and perform time synchronization and validity verification on the multi-source real-time operating data to form a standardized system status dataset;

[0007] S2: Based on the standardized system state dataset and meteorological forecast data, generate the power generation prediction curve of the photovoltaic unit and the power demand prediction curve of the charging load within the future preset time window;

[0008] S3: With the goal of minimizing the total system operating cost, a mixed integer linear programming model is constructed based on a standardized system state dataset, power generation prediction curve, and power demand prediction curve, which includes power balance constraints, equipment operation constraints, and user demand constraints.

[0009] S4: Solve the mixed integer linear programming model to obtain the optimal power allocation plan for each controllable unit within the future preset time window, and send the real-time reference power command corresponding to the current moment in the power allocation plan to the grid interaction interface, energy storage converter and bidirectional charging pile;

[0010] S5: After a preset optimization cycle, return to steps S1 to S4 to start a new round of optimization process with the updated system status data, so as to realize closed-loop rolling optimization control.

[0011] The present invention is further configured such that S1 includes:

[0012] Real-time operating parameters and status variables of the power supply side, grid side and load side are collected from photovoltaic inverters, energy storage systems, grid connection points, charging piles and load circuits in the station through the corresponding industrial communication protocols.

[0013] Outlier detection and removal are performed on the collected raw data stream, and interpolation is used to fill in missing data segments;

[0014] The cleaned data is timestamped based on a high-precision clock source and aggregated at preset optimization intervals to generate instantaneous or average system status values ​​with unified time labels.

[0015] The aggregated data is organized according to a preset structure to form a standard data frame containing timestamps, power of each unit, key state variables and external signals, which serves as a standardized system state dataset.

[0016] The present invention is further configured such that S2 includes:

[0017] Based on the obtained meteorological forecast data for the future preset time period, combined with the rated parameters and physical characteristics of photovoltaic modules, a deterministic physical model is used to calculate and generate the power generation prediction curve of the photovoltaic unit within the future preset time window;

[0018] Based on the statistical data of real-time charging demand and historical charging behavior of vehicles at the station in the standardized system state dataset, the future new charging load is predicted through a probabilistic model.

[0019] The deterministic charging demand of vehicles already at the station is superimposed with the probabilistic charging demand of future new vehicles, and the superposition result is statistically analyzed using the Monte Carlo simulation method to generate a power demand prediction curve for the total charging load within a preset time window.

[0020] The present invention is further configured such that S3 includes:

[0021] Define continuous decision variables related to the power interaction with the grid, the charging and discharging power of the energy storage system, and the charging and discharging power of each bidirectional charging pile; and define binary decision variables related to the charging and discharging states of the energy storage system and the bidirectional charging piles.

[0022] Construct an objective function that minimizes the total operating cost of the system, which includes the cost of purchasing electricity from the grid, the revenue from selling electricity to the grid, the aging cost of energy storage equipment, and the penalty cost for not meeting charging demand.

[0023] Establish a set of system constraints, which includes: system power balance constraints based on power prediction, physical operation constraints of energy storage system and bidirectional charging piles, and completion constraints to ensure the charging needs of electric vehicle users.

[0024] A mixed-integer linear programming model is constructed by integrating the decision variables, objective function, and constraint set.

[0025] The present invention is further configured such that S4 includes:

[0026] The mixed-integer linear programming model is converted into the standard input format of the mathematical programming solver, and the solution parameters are set. The solver is then called to perform numerical solutions to obtain the optimal solution set for the decision variables.

[0027] If the solution is successfully obtained, extract the values ​​of the decision variables corresponding to the first time interval within the optimization window from the optimal solution set;

[0028] The extracted decision variable values ​​are mapped to the real-time reference power commands of the grid interaction interface, energy storage converter, and each bidirectional charging pile.

[0029] Real-time reference power commands are sent to the corresponding device controllers to execute the power allocation plan.

[0030] The present invention is further configured such that S5 includes:

[0031] After issuing the real-time reference power command for the current cycle, wait for the preset optimization cycle and automatically trigger a new round of optimization process;

[0032] When starting a new round of optimization, the prediction curve is updated on a rolling basis based on the latest collected system status data, and the initial conditions of the mixed integer linear programming model are updated.

[0033] The optimization window is scrolled forward by one preset optimization cycle. Using the updated prediction curve, model initial conditions, and optimization window as input, S1 to S4 are executed repeatedly to form closed-loop feedback control and realize the dynamic operation optimization of the system.

[0034] The present invention is further configured such that the rolling update of the prediction curve includes: rolling the power generation prediction curve and the power demand prediction curve forward along the time axis by one preset optimization period, and discarding outdated data segments.

[0035] The present invention is further configured such that the issuance and execution of the real-time reference power command is achieved by the central controller deployed in the station communicating in real time with the power grid interface, energy storage converter and bidirectional charging pile through industrial Ethernet and using ModbusTCP or MQTT communication protocols.

[0036] The present invention is further configured such that the method includes: displaying the power generation prediction curve, the power demand prediction curve, and the real-time reference power command through a graphical interface in real time for monitoring the system operation status.

[0037] This invention also provides a dynamic operation optimization system for an integrated supercharging system encompassing power generation, grid, load, and energy storage, the system comprising:

[0038] Preprocessing module: Periodically acquires multi-source real-time operating data of the supercharging system, and performs time synchronization and validity verification on the multi-source real-time operating data to form a standardized system status dataset;

[0039] Ultra-short-term forecasting module: Based on standardized system state datasets and meteorological forecast data, it generates power generation forecast curves for photovoltaic units and power demand forecast curves for charging loads within a preset time window in the future;

[0040] Optimization Modeling Module: With the goal of minimizing the total system operating cost, a mixed integer linear programming model is constructed based on a standardized system state dataset, power generation prediction curve, and power demand prediction curve, which includes power balance constraints, equipment operation constraints, and user demand constraints.

[0041] The optimization control module solves the mixed-integer linear programming model to obtain the optimal power allocation plan for each controllable unit within the future preset time window, and sends the real-time reference power command corresponding to the current moment in the power allocation plan to the grid interaction interface, energy storage converter and bidirectional charging pile.

[0042] Rolling optimization module: After a preset optimization cycle, return to steps S1 to S4 to start a new round of optimization process with updated system status data, thereby realizing closed-loop rolling optimization control.

[0043] This invention provides a dynamic operation optimization method and system for an integrated supercharging system with grid-source, load, and energy storage. The method comprises: S1: periodically acquiring multi-source real-time operation data of the supercharging system and performing time synchronization and validity verification on the multi-source real-time operation data to form a standardized system state dataset; S2: generating power generation prediction curves for photovoltaic units and power demand prediction curves for charging loads within a future preset time window based on the standardized system state dataset and weather forecast data; S3: constructing a mixed-integer linear programming model containing power balance constraints, equipment operation constraints, and user demand constraints, based on the standardized system state dataset, power generation prediction curves, and power demand prediction curves, with the goal of minimizing the total system operating cost; S4: solving the mixed-integer linear programming model to obtain the optimal power allocation plan for each controllable unit within the future preset time window, and issuing the real-time reference power command corresponding to the current moment in the power allocation plan to the grid interaction interface, energy storage converter, and bidirectional charging piles; S5: after a preset optimization cycle, returning to steps S1 to S4 to start a new round of optimization with updated system state data, achieving closed-loop rolling optimization control. The beneficial effects include:

[0044] 1. By constructing an accurate mixed-integer linear programming model and implementing closed-loop rolling optimization, the economic efficiency of the supercharging system is maximized. This method not only comprehensively considers multiple factors such as grid time-of-use pricing, V2G reverse electricity sales revenue, energy storage battery aging costs, and user charging service quality, but also incorporates the uncertainty of photovoltaic output and charging demand into the optimization framework. This enables the dynamic generation of globally optimal power allocation strategies, effectively reducing the overall electricity cost of supercharging stations. Furthermore, through precise control of energy storage charging and discharging strategies and V2G timing, the local renewable energy absorption rate and asset utilization rate are improved.

[0045] 2. Transforming supercharging stations from traditional single-load electricity consumption into intelligent flexible nodes with proactive response capabilities enhances their friendliness to the distribution network. Through a rolling optimization mechanism based on model predictive control, they can smooth their own net load curve, effectively suppressing the impact of disorderly charging on the power grid. At the same time, they have the ability to participate in grid demand response and provide ancillary services such as peak shaving and valley filling. While achieving their own optimized operation, they can actively respond to grid dispatch instructions, providing important flexibility support for the stable and efficient operation of the distribution network, helping to alleviate expansion pressure and promoting the construction of smart grids.

[0046] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings:

[0048] Figure 1 A flowchart illustrating a dynamic operation optimization method for an integrated supercharging system with power generation, grid, load, and storage, as shown in an exemplary embodiment of the present invention;

[0049] Figure 2 This is a schematic diagram illustrating the structure of a dynamic operation optimization system for an integrated supercharging system with power generation, grid, load, and storage, as an exemplary embodiment of the present invention. Detailed Implementation

[0050] The embodiments of the present invention will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention and not for limiting the scope of protection of the present invention.

[0051] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0052] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0053] Example 1:

[0054] A dynamic operation optimization method for an integrated supercharging system with power generation, grid, load, and storage, such as... Figure 1 As shown, it includes:

[0055] S1: Periodically acquire multi-source real-time operating data of the supercharging system, and perform time synchronization and validity verification on the multi-source real-time operating data to form a standardized system status dataset;

[0056] S2: Based on the standardized system state dataset and meteorological forecast data, generate the power generation prediction curve of the photovoltaic unit and the power demand prediction curve of the charging load within the future preset time window;

[0057] S3: With the goal of minimizing the total system operating cost, a mixed integer linear programming model is constructed based on a standardized system state dataset, power generation prediction curve, and power demand prediction curve, which includes power balance constraints, equipment operation constraints, and user demand constraints.

[0058] S4: Solve the mixed integer linear programming model to obtain the optimal power allocation plan for each controllable unit within the future preset time window, and send the real-time reference power command corresponding to the current moment in the power allocation plan to the grid interaction interface, energy storage converter and bidirectional charging pile;

[0059] S5: After a preset optimization cycle, return to steps S1 to S4 to start a new round of optimization process with the updated system status data, so as to realize closed-loop rolling optimization control.

[0060] The present invention is further configured such that S1 includes:

[0061] Real-time operating parameters and status variables of the power supply side, grid side and load side are collected from photovoltaic inverters, energy storage systems, grid connection points, charging piles and load circuits in the station through the corresponding industrial communication protocols.

[0062] Outlier detection and removal are performed on the collected raw data stream, and interpolation is used to fill in missing data segments;

[0063] The cleaned data is timestamped based on a high-precision clock source and aggregated at preset optimization intervals to generate instantaneous or average system status values ​​with unified time labels.

[0064] The aggregated data is organized according to a preset structure to form standard data frames containing timestamps, power of each unit, key status variables, and external signals, serving as a standardized system status dataset. Specifically, this is accomplished through a communication management unit and various communication interface components deployed at the station control layer. Data interaction is performed using corresponding standard industrial communication protocols tailored to the different characteristics of equipment on the power supply side, grid side, and load side. For photovoltaic units, the communication management unit acts as a client, establishing a network connection with the photovoltaic inverter and periodically reading key parameters characterizing its operating status from predetermined register addresses within the inverter, including DC-side input power, AC-side output power, and internal temperature. Data acquisition for energy storage units is conducted via an independent bus communication link. It synchronously interacts with the battery management system and energy storage converter to obtain the comprehensive electrical status, remaining capacity information, health indicators, allowable power limits, and actual operating mode and power value of the battery cluster. Monitoring data on the grid side comes from smart metering devices installed at the grid connection point. Through request commands following their communication protocols, it periodically obtains active and reactive power data from the grid interaction point, as well as current electricity price information. Simultaneously, through a secure network interface service, it obtains potential demand response event commands from the grid dispatch platform in a polling manner. These commands contain structured information such as event identifier, control type, target power value, and effective time window. Load-side data is monitored in real-time by the charging pile management system to track the operating status and output power of each charging pile. The system collects data on charging rate and charged quantity, and obtains the battery status and charging needs of connected vehicles through the communication protocol between the vehicle and the charging pile. The charging goals and expected times set by users via mobile applications are persistently stored in the station's database and periodically queried by the data acquisition service. The power consumption of regular loads within the station is calculated by obtaining the current and voltage data of the circuit through intelligent electrical measurement elements, and then deriving the active power value according to the AC power calculation formula. The collected raw data stream undergoes a rigorous data cleaning process to ensure quality. A preset-length historical data window is maintained for continuously changing power data sequences, and statistical principles are used to identify outliers for each newly arrived data point, specifically using a judgment criterion based on the historical data mean and standard deviation. If the deviation from the historical data trend exceeds a preset threshold, it is judged as abnormal and removed, and filled using an interpolation strategy based on the effective values ​​before and after. For data loss caused by communication interruption, the same interpolation algorithm is used for repair, and an alarm is triggered when the loss continues to exceed the limit. To ensure time series consistency, all data points are marked with a precise timestamp provided by a high-precision clock source during acquisition, and the clock is kept synchronized with the time server through a standard time synchronization protocol to ensure the uniformity of the time base of the entire station's data. A preset time interval is used as the optimization period, and at the end of each optimization period, all time series data acquired within the optimization period are aggregated. For state parameters, the instantaneous value at the end of the optimization period is taken as representative.For flow parameters such as power, the average of all valid sampled values ​​within the optimization period is calculated by integration, thereby unifying all high-frequency collected raw data to the optimization time granularity. The aggregated data is assembled into a standardized data frame structure according to predefined specifications. This data frame serves as a snapshot of the system's state at the end of a certain optimization period, and its fields include key information such as timestamp, photovoltaic output, energy storage status and power, grid interaction power, total electric vehicle load, station base load, and current electricity price.

[0065] The present invention is further configured such that S2 includes:

[0066] Based on the obtained meteorological forecast data for the future preset time period, combined with the rated parameters and physical characteristics of photovoltaic modules, a deterministic physical model is used to calculate and generate the power generation prediction curve of the photovoltaic unit within the future preset time window;

[0067] Based on the statistical data of real-time charging demand and historical charging behavior of vehicles at the station in the standardized system state dataset, the future new charging load is predicted through a probabilistic model.

[0068] The deterministic charging demand of vehicles already at the station is superimposed with the probabilistic charging demand of future new vehicles. Monte Carlo simulation is then used to statistically analyze the superposition results, generating a power demand forecast curve for the total charging load within a preset future time window. Specifically, the power generation forecast of the photovoltaic unit first obtains hourly total solar irradiance and ambient temperature forecasts for the area where the supercharging station is located by calling a meteorological data service interface. A power calculation model is then established based on the physical characteristics of the photovoltaic modules. The core basis of this model is the linear relationship between photoelectric conversion efficiency and irradiance intensity, as well as the temperature effect of semiconductor materials. Specifically, this includes: firstly, comparing the forecasted irradiance with standard measured... The relative irradiance ratio is calculated by comparing the baseline irradiance defined under the test conditions. This ratio is then multiplied by the rated peak power of the photovoltaic array to obtain the theoretical output power under standard test temperature conditions. Subsequently, temperature characteristic correction is performed. The operating temperature of the photovoltaic panel is calculated based on the predicted ambient temperature and the thermal characteristic parameters of the components. The deviation from the standard test temperature is calculated, and this deviation is substituted into the negative correlation parameter characterizing the temperature coefficient of the component's output power to obtain the power temperature correction factor. Finally, the theoretical output power is multiplied by the power temperature correction factor to obtain the predicted photovoltaic power generation value after dual calibration of irradiance and temperature. This process is iteratively performed on all consecutive time nodes within the prediction time window. The calculation process ultimately generates a power generation prediction curve for photovoltaic units with time continuity. The specific implementation process of power demand prediction for total charging load is based on a comprehensive analysis of historical behavior patterns and real-time status monitoring. Its input data sources include two components: first, real-time session information of all currently charging vehicles at the station, parsed from a standardized system status dataset, specifically including the real-time remaining battery percentage of each electric vehicle, the target charging percentage set by the user through the interactive interface, and the planned departure timestamp; second, long-term charging behavior records extracted from the historical operation database. This structured data includes the start time and initial battery status of each charging event within the historical period. The system considers parameters such as the charging status, termination charge status, duration, and charging power level used, and performs cluster analysis based on date type and time period characteristics. The prediction calculation adopts a hybrid modeling method combining determinism and probabilistics. The deterministic part is for vehicles currently on-site. It calculates the difference between the target charge and the current charge, and combines it with the battery's rated capacity to obtain the total energy to be charged. Then, it calculates the average charging power required based on the time difference between the current time and the planned departure time. The battery's rated capacity needs to be determined according to the technical parameters corresponding to the vehicle model. The average power calculation needs to consider the power limitations of the charging equipment and the maximum acceptable charging rate of the battery. This part of the basic charging load has clear time boundaries and energy requirements.The probabilistic part, addressing the prediction of future new charging load, firstly establishes a vehicle arrival rate model that varies over time based on statistical analysis of historical data from the same period. A non-homogeneous stochastic process considering time-period characteristics is used to describe the arrival patterns, including the distinction between weekdays and holidays, and the periodicity of morning and evening peak hours. Simultaneously, by fitting the probability distribution of historical data, a multidimensional joint probability model of the initial and target battery levels of new vehicles is established. This multidimensional joint probability model needs to undergo statistical testing to ensure its consistency with the actual data distribution characteristics. Subsequently, statistical simulation methods are used for multi-scenario simulations, specifically employing Monte Carlo stochastic simulation technology. In each simulation, vehicles are randomly generated based on the arrival rate model. The system calculates the arrival time series of vehicles and configures charging demand parameters for each new vehicle according to a joint probability distribution. Each simulation must adhere to the physical constraints of the charging process, including upper and lower limits of charging power and battery capacity limitations. A complete charging load curve is formed by superimposing deterministic base charging loads and random new charging loads. Through statistical simulations using numerous independent random trials, the number of simulations must be determined according to the central limit theorem to ensure the convergence and stability of the statistical results. The mean value and probability interval estimation are performed on the load values ​​at each time point. The probability interval estimation uses quantile calculation methods, typically taking the upper and lower quartiles as representations of the uncertainty range. Finally, a power prediction curve for the total charging load with probability distribution characteristics is generated.

[0069] The present invention is further configured such that S3 includes:

[0070] Define continuous decision variables related to the power interaction with the grid, the charging and discharging power of the energy storage system, and the charging and discharging power of each bidirectional charging pile; and define binary decision variables related to the charging and discharging states of the energy storage system and the bidirectional charging piles.

[0071] Construct an objective function that minimizes the total operating cost of the system, which includes the cost of purchasing electricity from the grid, the revenue from selling electricity to the grid, the aging cost of energy storage equipment, and the penalty cost for not meeting charging demand.

[0072] A set of system constraints is established, including: system power balance constraints based on power prediction, physical operation constraints of the energy storage system and bidirectional charging piles, and completion constraints to ensure the charging needs of electric vehicle users. A mixed-integer linear programming model is constructed by integrating decision variables, the objective function, and the set of constraints. Specifically, the decision variables are first defined: grid interaction power needs to be defined as two independent non-negative continuous variables, representing the positive power purchased from the grid and the negative power fed back to the grid, respectively; the energy storage system needs to define its charging power and discharging power as two non-negative continuous variables, and a pair of mutually exclusive binary state variables need to be introduced to represent its state of charging, discharging, or waiting. The machine's operating status; for each bidirectional charging pile with V2G functionality, two non-negative continuous variables, charging power and discharging power, need to be defined, along with corresponding binary state variables to describe its operating mode; after clarifying the decision variables, an objective function needs to be constructed with minimizing the total system operating cost as the single objective. The total operating cost includes the following components: the electricity purchase cost calculated based on the grid purchase power and real-time electricity price; the negative cost calculated based on the electricity sold to the grid and the electricity price; the equipment depreciation cost calculated based on the energy storage system's charging and discharging power and unit aging coefficient; and the penalty cost incurred due to the actual delivered charging volume not reaching the planned charging volume; among which, the penalty cost needs to be calculated by calculating the planned charging volume. The difference between the available charge and the actual deliverable charge is multiplied by a preset penalty weighting coefficient to determine the system's capacity. Finally, a set of system constraints needs to be established, with the core constraint being the real-time power balance equation for the supercharging system. This equation requires that the sum of the photovoltaic predicted power, the net grid interaction power, and the net energy storage system discharge power equal the sum of the base load predicted power and the net charging power of all charging piles. The photovoltaic predicted power, the net grid interaction power, and the net energy storage system discharge power are obtained from the power generation prediction curve and the power demand prediction curve. The energy storage system must meet the following operational constraints: first, a mutual exclusion constraint between charging and discharging states, ensuring that it cannot be in both charging and discharging states simultaneously; second, upper and lower power limits, where the charging and discharging power values ​​are less than or equal to the rated limits. The system requires logical coupling through binary state variables; thirdly, dynamic state of charge (SCC) update constraints, calculating the SCC value for the next time period based on the charging and discharging power of the current time period; and SCC safety boundary constraints, requiring its value to always remain within a preset allowable range. For bidirectional charging piles, in addition to meeting operational constraints similar to those of energy storage systems, user demand constraints are also required to ensure that the battery SCC of the vehicle at the planned departure time is not lower than the preset target value. All constraints must be established independently based on each discrete time interval within the prediction time window. Through the above steps, the complex operational optimization problem is formalized into a standard mixed-integer linear programming model, providing a clear mathematical description for the subsequent solution algorithm.

[0073] The present invention is further configured such that S4 includes:

[0074] The mixed-integer linear programming model is converted into the standard input format of the mathematical programming solver, and the solution parameters are set. The solver is then called to perform numerical solutions to obtain the optimal solution set for the decision variables.

[0075] If the solution is successfully obtained, extract the values ​​of the decision variables corresponding to the first time interval within the optimization window from the optimal solution set;

[0076] The extracted decision variable values ​​are mapped to the real-time reference power commands of the grid interaction interface, energy storage converter, and each bidirectional charging pile.

[0077] The real-time reference power command is sent to the corresponding device controller to execute the power allocation plan. Further, the issuance and execution of the real-time reference power command are achieved by a central controller deployed within the station communicating in real-time with the power grid interface, energy storage converter, and bidirectional charging pile via an industrial Ethernet network and using Modbus TCP or MQTT communication protocols. Specifically, the objective function and constraint set contained in the defined mixed-integer linear programming model are instantiated into a complete mathematical model object according to the application programming interface specification of the selected commercial mathematical programming solver. This model object encapsulates all decision variables, the mathematical expression of the objective function, and the mathematical expressions of the constraints in the mixed-integer linear programming model. Subsequently, the key calculation parameters of the solver are configured, including the maximum allowable solution time threshold and the optimality gap tolerance threshold. After completing the calculation parameter configuration, the solver is invoked. The numerical optimization algorithm execution function initiates the numerical solution process for the mixed-integer linear programming model. This process involves the solver automatically invoking its embedded optimization algorithms, such as branch and bound, cutting plane, or a hybrid strategy, based on the configured mathematical model and calculation parameters. Through iterative calculation, it searches for the optimal solution set of decision variables that minimizes the objective function value under all constraints. After the solution process is complete, the solver returns the calculation results containing a solution status identifier and the corresponding set of decision variable values. The solution status identifier includes optimal solution, feasible solution, and no solution. When the status identifier is optimal solution, the optimal solution set of all decision variables is extracted from the calculation results. When the solution status identifier is optimal solution, the value corresponding to the first time interval within the optimization time window is extracted from the optimal solution set of the decision variables. The real-time reference power command at the grid interaction point is defined as the algebraic difference between the grid's purchased power value and the grid's sold power value at that moment. The real-time reference power command of the energy storage system is generated based on its charging power and discharging power values, where the charging power command is represented by a negative value and the discharging power command by a positive value.For each bidirectional charging pile, the real-time reference power command is defined as the algebraic difference between the charging power value and the discharging power value of that pile. A positive value represents the charging operating mode, and a negative value represents the discharging operating mode. The above real-time reference power command is mapped to the power control parameters of specific controllable devices. The real-time reference power command of the grid interaction point is configured as the active power setpoint of the grid interface converter; the real-time reference power command of the energy storage system is configured as the charging and discharging power setpoint of the energy storage converter; the real-time reference power command of each bidirectional charging pile is configured as the output power setpoint of the corresponding charging pile controller. The issuance and execution of the real-time reference power command are realized through the central controller deployed in the station. This controller establishes a real-time communication connection with the grid interaction interface device, energy storage converter, and bidirectional charging pile based on the industrial Ethernet network and using the Modbus TCP or MQTT communication protocol. By converting the real-time reference power command value generated by the optimized calculation into a communication message that can be recognized by each controlled device, the message is then transmitted to each device controller in real time through the industrial Ethernet. The device controller completes the message parsing and immediately executes the corresponding power setting, thereby completing the execution of the power allocation plan.

[0078] The present invention is further configured such that S5 includes:

[0079] After issuing the real-time reference power command for the current cycle, wait for the preset optimization cycle and automatically trigger a new round of optimization process;

[0080] When starting a new round of optimization, the prediction curve is updated on a rolling basis based on the latest collected system status data, and the initial conditions of the mixed integer linear programming model are updated.

[0081] The optimization window is rolled forward by one preset optimization cycle. Using the updated prediction curve, model initial conditions, and optimization window as input, S1 to S4 are repeated to form a closed-loop feedback control, thereby achieving dynamic operation optimization of the system. The invention is further configured such that the rolling update of the prediction curve includes: rolling the power generation prediction curve and the power demand prediction curve forward along the time axis by one preset optimization cycle, and discarding outdated data segments. Specifically, upon completion of the real-time reference power command issuance event for the current optimization cycle, a high-precision timer is started to begin timing, waiting for a time interval equal to the preset optimization cycle. When the timer reaches the set time limit, a trigger signal is automatically generated, and a new round of complete optimization process is initiated. When initiating a new round of optimization, the system status data is first updated. Based on the latest collected system status data, the initial value of the energy storage system's state of charge is reset, prioritizing the use of the latest measured state of charge value. As an initial condition, if the measured value is unavailable due to communication failure or other reasons, the state of charge value of the second time interval predicted in the previous cycle is used as the backup initial value. At the same time, the power generation prediction curve and the power demand prediction curve are updated in a rolling manner, and the two prediction curves are rolled forward along the time axis by one optimization cycle, and the data points of the earliest time segment that are outdated are discarded. Then, the optimization time window is rolled forward by one optimization cycle, so that the new optimization window covers the continuous time interval from the current moment to the future preset time range. The updated prediction curve, the updated model initial conditions, and the new optimization time window are used as inputs to re-execute the complete process from S1 to S4. Through the continuous operation of the periodic triggering, state update and optimization execution mechanism, a closed-loop control system with feedback correction capability is constructed, which can continuously optimize based on the latest operating status and external environment information to realize the dynamic operation optimization of the supercharging system.

[0082] The invention is further configured such that the method includes: displaying the power generation prediction curve, power demand prediction curve, and real-time reference power command in a graphical interface in real time for monitoring system operation status; specifically, the graphical interface synchronously displays a comparative analysis of the photovoltaic power generation prediction curve and the actual power output value, as well as a comparative analysis of the charging total load power demand prediction curve and the actual load value, through trend graphs; simultaneously, the interface displays the reference real-time power command and its execution status issued to the grid interaction interface, energy storage converter, and each bidirectional charging pile in real time through a power flow graph; in addition, the interface also provides historical data backtracking query and alarm event recording functions to ensure that operators can fully grasp the system operation status.

[0083] Example 2:

[0084] Please see Figure 2The exemplary dynamic operation optimization system for an integrated supercharging system comprising source, grid, load, and storage system includes:

[0085] Preprocessing module: Periodically acquires multi-source real-time operating data of the supercharging system, and performs time synchronization and validity verification on the multi-source real-time operating data to form a standardized system status dataset;

[0086] Ultra-short-term forecasting module: Based on standardized system state datasets and meteorological forecast data, it generates power generation forecast curves for photovoltaic units and power demand forecast curves for charging loads within a preset time window in the future;

[0087] Optimization Modeling Module: With the goal of minimizing the total system operating cost, a mixed integer linear programming model is constructed based on a standardized system state dataset, power generation prediction curve, and power demand prediction curve, which includes power balance constraints, equipment operation constraints, and user demand constraints.

[0088] The optimization control module solves the mixed-integer linear programming model to obtain the optimal power allocation plan for each controllable unit within the future preset time window, and sends the real-time reference power command corresponding to the current moment in the power allocation plan to the grid interaction interface, energy storage converter and bidirectional charging pile.

[0089] Rolling optimization module: After a preset optimization cycle, return to steps S1 to S4 to start a new round of optimization process with updated system status data, thereby realizing closed-loop rolling optimization control.

[0090] It should be noted that the dynamic operation optimization system for an integrated supercharging system with power generation, grid, load, and storage provided in the above embodiments and the dynamic operation optimization method for such a system are based on the same concept. The specific methods by which each module and unit performs its operations have been described in detail in the method embodiments and will not be repeated here. In practical applications, the dynamic operation optimization system for an integrated supercharging system with power generation, grid, load, and storage provided in the above embodiments can be configured to distribute the aforementioned functions among different functional modules as needed. That is, the internal structure of the system can be divided into different functional modules to complete all or part of the functions described above, and this is not a limitation here.

[0091] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A dynamic operation optimization method for an integrated supercharging system combining power generation, grid, load, and storage, characterized in that, include: S1: Periodically acquire multi-source real-time operating data of the supercharging system, and perform time synchronization and validity verification on the multi-source real-time operating data to form a standardized system status dataset; S2: Based on the standardized system state dataset and meteorological forecast data, generate the power generation prediction curve of the photovoltaic unit and the power demand prediction curve of the charging load within the future preset time window; S3: With the goal of minimizing the total system operating cost, a mixed integer linear programming model is constructed based on a standardized system state dataset, power generation prediction curve, and power demand prediction curve, which includes power balance constraints, equipment operation constraints, and user demand constraints. S4: Solve the mixed integer linear programming model to obtain the optimal power allocation plan for each controllable unit within the future preset time window, and send the real-time reference power command corresponding to the current moment in the power allocation plan to the grid interaction interface, energy storage converter and bidirectional charging pile. S5: After a preset optimization cycle, return to steps S1 to S4 to start a new round of optimization process with the updated system status data, so as to realize closed-loop rolling optimization control.

2. The dynamic operation optimization method for an integrated supercharging system based on source, grid, load, and storage as described in claim 1, characterized in that, S1 includes: Real-time operating parameters and status variables of the power supply side, grid side and load side are collected from photovoltaic inverters, energy storage systems, grid connection points, charging piles and load circuits in the station through the corresponding industrial communication protocols. Outlier detection and removal are performed on the collected raw data stream, and interpolation is used to fill in missing data segments; The cleaned data is timestamped based on a high-precision clock source and aggregated at preset optimization intervals to generate instantaneous or average system status values ​​with unified time labels. The aggregated data is organized according to a preset structure to form a standard data frame containing timestamps, power of each unit, key state variables and external signals, which serves as a standardized system state dataset.

3. The dynamic operation optimization method for an integrated supercharging system based on source, grid, load, and storage as described in claim 1, characterized in that, S2 includes: Based on the obtained meteorological forecast data for the future preset time period, combined with the rated parameters and physical characteristics of photovoltaic modules, a deterministic physical model is used to calculate and generate the power generation prediction curve of the photovoltaic unit within the future preset time window; Based on the statistical data of real-time charging demand and historical charging behavior of vehicles at the station in the standardized system state dataset, the future new charging load is predicted through a probabilistic model. The deterministic charging demand of vehicles already at the station is superimposed with the probabilistic charging demand of future new vehicles, and the superposition result is statistically analyzed using the Monte Carlo simulation method to generate a power demand prediction curve for the total charging load within a preset time window.

4. The dynamic operation optimization method for an integrated supercharging system based on source, grid, load, and storage as described in claim 1, characterized in that, S3 includes: Define continuous decision variables related to the power interaction with the grid, the charging and discharging power of the energy storage system, and the charging and discharging power of each bidirectional charging pile; and define binary decision variables related to the charging and discharging states of the energy storage system and the bidirectional charging piles. Construct an objective function that minimizes the total operating cost of the system, which includes the cost of purchasing electricity from the grid, the revenue from selling electricity to the grid, the aging cost of energy storage equipment, and the penalty cost for not meeting charging demand. Establish a set of system constraints, which includes: system power balance constraints based on power prediction, physical operation constraints of energy storage system and bidirectional charging piles, and completion constraints to ensure the charging needs of electric vehicle users. A mixed-integer linear programming model is constructed by integrating the decision variables, objective function, and constraint set.

5. The dynamic operation optimization method for an integrated supercharging system based on source, grid, load, and storage as described in claim 1, characterized in that, S4 includes: The mixed-integer linear programming model is converted into the standard input format of the mathematical programming solver, and the solution parameters are set. The solver is then called to perform numerical solutions to obtain the optimal solution set for the decision variables. If the solution is successfully obtained, extract the values ​​of the decision variables corresponding to the first time interval within the optimization window from the optimal solution set; The extracted decision variable values ​​are mapped to the real-time reference power commands of the grid interaction interface, energy storage converter, and each bidirectional charging pile. Real-time reference power commands are sent to the corresponding device controllers to execute the power allocation plan.

6. The dynamic operation optimization method for an integrated supercharging system based on source, grid, load, and storage as described in claim 1, characterized in that, S5 includes: After issuing the real-time reference power command for the current cycle, wait for the preset optimization cycle and automatically trigger a new round of optimization process; When starting a new round of optimization, the prediction curve is updated on a rolling basis based on the latest collected system status data, and the initial conditions of the mixed integer linear programming model are updated. The optimization window is scrolled forward by one preset optimization cycle. Using the updated prediction curve, model initial conditions, and optimization window as input, S1 to S4 are executed repeatedly to form closed-loop feedback control and realize the dynamic operation optimization of the system.

7. The dynamic operation optimization method for an integrated supercharging system based on source, grid, load, and storage as described in claim 6, characterized in that, The rolling update of the prediction curves includes: rolling the power generation prediction curve and the power demand prediction curve forward along the time axis by one preset optimization cycle, and discarding outdated data segments.

8. The dynamic operation optimization method for an integrated supercharging system based on power generation, grid, load, and storage as described in claim 1, characterized in that, The issuance and execution of the real-time reference power command are achieved by the central controller deployed in the station through industrial Ethernet and by using ModbusTCP or MQTT communication protocols to communicate in real time with the power grid interface, energy storage converter and bidirectional charging pile.

9. The dynamic operation optimization method for an integrated supercharging system based on source-grid-load-storage as described in claim 1, characterized in that, The method also includes: displaying the power generation prediction curve, power demand prediction curve, and real-time reference power command in a graphical interface in real time for monitoring the system's operating status.

10. A dynamic operation optimization system for an integrated supercharging system of power generation, grid, load, and storage, used to implement the dynamic operation optimization method for an integrated supercharging system of power generation, grid, load, and storage as described in any one of claims 1-9, characterized in that, include: Preprocessing module: Periodically acquires multi-source real-time operating data of the supercharging system, and performs time synchronization and validity verification on the multi-source real-time operating data to form a standardized system status dataset; Ultra-short-term forecasting module: Based on standardized system state datasets and meteorological forecast data, it generates power generation forecast curves for photovoltaic units and power demand forecast curves for charging loads within a preset time window in the future; Optimization Modeling Module: With the goal of minimizing the total system operating cost, a mixed integer linear programming model is constructed based on a standardized system state dataset, power generation prediction curve, and power demand prediction curve, which includes power balance constraints, equipment operation constraints, and user demand constraints. The optimization control module solves the mixed-integer linear programming model to obtain the optimal power allocation plan for each controllable unit within the future preset time window, and sends the real-time reference power command corresponding to the current moment in the power allocation plan to the grid interaction interface, energy storage converter and bidirectional charging pile. Rolling optimization module: After a preset optimization cycle, return to steps S1 to S4 to start a new round of optimization process with updated system status data, thereby realizing closed-loop rolling optimization control.

Images