Large-scale electric vehicle charging and discharging rate multi-level optimization regulation and control method

Abstract

The invention relates to the technical field of electric vehicle charging and discharging regulation and control, in particular to a large-scale electric vehicle charging and discharging rate multi-level optimization regulation and control method. Establishing a daily driving mileage prediction model by analyzing historical driving and charging data of the electric vehicle, and calculating a charging demand of the daily driving mileage according to the daily driving mileage prediction model; designing an optimization model based on linear programming based on the determined charging demand, and obtaining reasonable charging power of each electric vehicle in different time periods; on the basis of the determined charging demand, constructing a multi-level optimization model; under the guidance of the constructed multi-objective optimization model, a specific charging and discharging strategy of each electric vehicle is refined by adopting a heuristic algorithm; the charging power is dynamically adjusted by combining real-time data of the electric vehicle and the power grid; and based on the dynamic adjustment result, the charging power of different levels is coordinated, and multi-level cooperative control is realized. According to the large-scale electric vehicle charging and discharging rate multi-level optimization regulation and control method, efficient and stable operation of a charging system is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric vehicle charging and discharging control, and in particular to a multi-level optimization control method for charging and discharging rates of large-scale electric vehicles. Background Art

[0002] With the rapid development of the electric vehicle market, the demand for charging of large-scale electric vehicles is growing. However, the disorderly charging behavior of large-scale electric vehicles may bring a series of problems to the power grid, such as grid load fluctuations and reduced power quality. These problems will not only affect the stable operation of the power grid, but may also lead to adverse consequences such as increased charging costs and reduced user charging experience.

[0003] Traditional charging control methods usually only focus on the charging situation of a single charging station or a local area, and lack global consideration of the entire charging network. This local optimization method is difficult to maximize the overall efficiency of the charging system, and it is also difficult to effectively balance the load on the power grid. In addition, traditional methods are often not flexible enough to deal with the dynamic changes in the charging demand of electric vehicles, and it is difficult to adjust the charging strategy in real time to adapt to changes in actual conditions. In order to solve these problems, a multi-level optimization and control method is needed that can be coordinated from multiple levels such as a single electric vehicle, charging station, region, and the entire charging network. By comprehensively considering the charging needs of electric vehicles, the load conditions of the power grid, the capacity of charging facilities, and various external factors (such as weather, holidays, etc.), the reasonable allocation and optimal scheduling of charging power can be achieved, thereby improving the overall efficiency and reliability of the charging system, ensuring the stable operation of the power grid, while reducing charging costs and improving user satisfaction.

[0004] In summary, it is of great practical significance and urgency to develop a multi-level optimization and control method for large-scale electric vehicle charging and discharging rates. Summary of the invention

[0005] In view of the above deficiencies in the prior art, the purpose of the present invention is to provide a multi-level optimization and control method for charging and discharging rates of large-scale electric vehicles to achieve efficient and stable operation of the charging system.

[0006] In order to achieve the above objectives, the present invention adopts the following technical solutions:

[0007] The multi-level optimization control method of charging and discharging rate of large-scale electric vehicles includes the following steps:

[0008] S1. Modeling of electric vehicle charging and discharging loads: by analyzing the historical driving and charging data of electric vehicles, a daily mileage prediction model is established to predict the daily mileage of electric vehicles, and the charging demand of the daily mileage is calculated according to the daily mileage prediction model;

[0009] S2. Determine the charging and discharging power. Based on the charging demand determined in S1, design an optimization model based on linear programming to obtain the reasonable charging power of each electric vehicle in different time periods; achieve fair distribution of power capacity, ensure the power utilization efficiency of the charging station and the reasonable attenuation of the battery charging rate;

[0010] S3. Multi-level optimization target construction, based on the charging demand determined in S2, a multi-level optimization model is constructed;

[0011] S4. Optimization of charging and discharging strategies. Under the guidance of the multi-objective optimization model constructed in S3, a heuristic algorithm is used to refine the specific charging and discharging strategies of each electric vehicle;

[0012] S5. Real-time dispatching and control, dynamically adjusting the charging power according to the charging and discharging strategy determined in S4 and combining the real-time data of electric vehicles and power grids;

[0013] S6. Multi-level collaborative control, based on the dynamic adjustment of charging power results in S5, coordinates the charging power at different levels, realizes multi-level collaborative control, and ensures multi-level collaboration from a single charging station to the entire charging network.

[0014] Furthermore, in S1, the method for establishing the daily mileage prediction model is: a method combining time series analysis and machine (random forest model) is used to fuse the prediction results of the two models by weighted average method to obtain the final daily mileage prediction value:

[0015] D pr =ω1×D ar +ω2×D rf ;

[0016]

[0017] D rf = f[X];

[0018] Among them, ω1 and ω2 are weights; μ is a constant term, is the autoregressive coefficient, which reflects the influence of the past daily mileage on the current prediction value; p is the autoregressive order, which means that the daily mileage data at several past time points are considered; D t-1 is the daily mileage at time i in the past; ε t is the white noise error term; θ j is the moving average coefficient, q is the moving average order; f is the random forest model function; X is the input vector containing features such as holidays and weather.

[0019] Furthermore, in S1, the charging demand determined by the daily mileage = daily mileage × energy consumption per unit mileage; the energy consumption per unit mileage is determined according to the specific model and historical data of the electric vehicle.

[0020] Further, the S2 includes:

[0021] S21. Goal setting, including fairness indicators, power efficiency indicators, and reasonable decay battery charging rates;

[0022] Among them, the goal of the fairness index is to minimize the difference between the charging powers of different electric vehicles to achieve fair distribution. The fairness index is defined as:

[0023]

[0024] F represents the fairness index, P sch,i represents the dispatchable power of the i-th electric vehicle; represents the average dispatchable power of all electric vehicles;

[0025] The goal of the power usage efficiency index is to maximize the power usage efficiency of the charging station. The power usage efficiency index is defined as:

[0026]

[0027] P i,j represents the charging power of the i-th electric vehicle in the j-th time period; Δt represents the length of the time period; P total,j represents the total power capacity of the charging station in the jth time period; m is the total number of time periods; n is the total number of electric vehicles;

[0028] The goal of a reasonable decay battery charging rate is to determine the power upper limit of different charging stages according to the battery's state of charge (SOC). The relationship between the battery's charging power and SOC is expressed by the following function:

[0029]

[0030] P ch (SOC i,j ) represents the maximum charging power of the i-th electric vehicle at a specific SOC in the j-th time period; P max is the maximum charging power allowed by the battery; SOC i,j is the battery state of charge of the i-th electric vehicle in the j-th time period; SOC max is the maximum state of charge of the battery; α is an adjustment parameter used to control the speed at which the charging power decreases as SOC increases;

[0031] S22. Design an optimization model based on linear programming:

[0032]

[0033] Among them, Psch represents the dispatchable power of a single electric vehicle; P max is the maximum charging power, which limits the upper limit of the charging power of each electric vehicle; P total is the total power of the charging station, reflecting the total power that the charging station can provide; n is the total number of electric vehicles; m is the total number of time periods; P i is the power allocated to the i-th electric vehicle; is an indicator function, which is 1 when the i-th electric vehicle is charging in the j-th time period, otherwise it is 0, indicating that the impact of the charging state of each electric vehicle in different time periods on the power allocation is taken into account;

[0034] S23. Determine the constraints:

[0035] (1) The sum of the charging power of all electric vehicles cannot exceed the total power capacity of the charging station, that is:

[0036]

[0037] P i,j represents the charging power of the i-th electric vehicle in the j-th time period; P total,j represents the total power capacity of the charging station in the jth time period; n is the total number of electric vehicles; m is the total number of time periods;

[0038] (2) The charging power of each electric vehicle cannot exceed its maximum allowable value, that is:

[0039] 0≤P i，j ≤P i，max ;

[0040] P i,j represents the charging power of the i-th electric vehicle in the j-th time period; P i，max is the maximum charging power of the i-th electric vehicle;

[0041] (3) Consider the power upper limit at different charging stages to avoid damage to battery life caused by overcharging. According to the charging characteristics of the battery, the charging power should meet the following requirements:

[0042] P i，j ≤P ch (SOC i,j );

[0043] P ch (SOC i,j ) represents the maximum charging power of the i-th electric vehicle at a specific SOC in the j-th time period; P i,j represents the charging power of the i-th electric vehicle in the j-th time period;

[0044] S24. Solve the model using a linear programming solver, input the objective function obtained in S21 and the constraints in S22 into the solver to obtain the reasonable charging power of each electric vehicle in different time periods.

[0045] Further, the S3 includes:

[0046] S31. Determine optimization objectives, including operating cost objectives, user satisfaction objectives, and grid load balance objectives;

[0047] Among them, the cost function is expressed as:

[0048]

[0049] Where C(Q) represents the functional relationship between operating cost and charging demand; Q i represents the charging demand of the i-th electric vehicle; n represents the total number of electric vehicles; χ is the electricity procurement cost coefficient per unit charging demand;

[0050] The user satisfaction function is expressed as:

[0051]

[0052] Among them, U(Q) represents the functional relationship between user satisfaction and charging demand; ω represents the adjustment coefficient, P i represents the charging power of the i-th electric vehicle; Q i represents the charging demand of the i-th electric vehicle;

[0053] The grid load balancing objective is expressed as:

[0054]

[0055] Where B(Q) represents the functional relationship between grid load balance and charging demand; η is the adjustment coefficient; n represents the total number of electric vehicles; P i represents the charging power of the i-th electric vehicle; Q i represents the charging demand of the i-th electric vehicle; L max is the maximum value of grid load balance;

[0056] S32. Multi-level optimization objective function construction, integrating the optimization objectives determined in S31 into a model, constructing multi-level optimization objectives, and the multi-level optimization objective function is expressed as:

[0057] G(Q)=αC(Q)+βU(Q)+γB(Q);

[0058] G(Q) represents the multi-level optimization objective function; α, β, γ are weight coefficients used to adjust the importance of each objective; C(Q), U(Q), B(Q) represent the operating cost objective, user satisfaction objective, and grid load balance objective, respectively;

[0059] Objective function constraints:

[0060] Charging power constraints, Q total Is the charging demand online; Q i represents the charging demand of the i-th electric vehicle; n is the total number of electric vehicles.

[0061] Further, the S4 includes:

[0062] Set the initial parameters of the algorithm, such as population size, number of iterations, etc.

[0063] Representing the charging and discharging strategy of each electric vehicle as a coding form, such as a binary code or a real number code;

[0064] Randomly generate a set of initial charging and discharging strategies as the initial population of the algorithm;

[0065] Calculate the fitness value according to the multi-objective optimization function constructed by S3;

[0066] Among them, the fitness function is expressed as:

[0067]

[0068] In the formula, F P Fitness value; C max is the maximum value of operating cost; U max is the maximum value of user satisfaction; L max is the maximum value of the grid load balance; C(Q), U(Q), and L(Q) represent the operating cost target, user satisfaction target, and grid load balance target, respectively; ω C is the target weight of operating cost; U is the target weight of user satisfaction; ω B is the target weight for grid load balancing;

[0069] A set of initial charging and discharging strategies are randomly generated, each strategy is represented as a vector, which contains the charging power or discharging power of each electric vehicle in each time period; through the selection operation, excellent individuals are selected to enter the next generation according to the value of the fitness function; through the crossover operation, the genes of the selected individuals are exchanged to generate new individuals; through the mutation operation, some genes of the individuals are randomly changed to increase the diversity of the population;

[0070] Repeat the above steps until the termination condition is met, such as reaching the maximum number of iterations or finding a satisfactory solution.

[0071] Further, the S5 includes:

[0072] S51. Collect relevant data of electric vehicles and power grids in real time through sensors, communication networks and other means, and transmit the real-time data to the control center; the real-time data includes parameters such as electric vehicle current, voltage, battery temperature, and frequency, voltage, power and other information of the power grid;

[0073] S52. According to the charging and discharging strategy determined in S4, understand the ideal charging power and time arrangement of each electric vehicle under different conditions;

[0074] S53. Analyze the data collected in real time to determine whether the actual demand of the current electric vehicle and the load of the power grid are consistent with the charging and discharging strategy of S4; if there is a difference, adjustments need to be made;

[0075] S54. Based on the analysis results of real-time data and combined with the charging and discharging strategy of S4, a real-time scheduling algorithm is used to dynamically adjust the charging power;

[0076] S55. Send the adjusted charging power instruction to the charging pile or charging equipment to ensure that they can charge according to the new power;

[0077] S56. During the charging process, the real-time data of the electric vehicle and the power grid are continuously monitored, and the charging power is continuously adjusted according to the actual situation to ensure the real-time, adaptability and stability of the charging process.

[0078] Furthermore, in S54, the real-time scheduling algorithm is expressed as:

[0079]

[0080] P c ' ,t represents the value of the adjusted charging power at time t; P c,t represents the current charging power value at time t; λ is the charging power adjustment coefficient; sign(x) is the sign function, when x>0, sign(x)=1; when x=0, sign(x)=0; when x<0, sign(x)=-1; represents the ideal charging power value at time t determined according to the charging and discharging strategy in S4; Indicates taking and ΔP max The smaller value of ΔP max is the preset maximum adjustment of charging power; μ is the grid load adjustment coefficient; represents the actual load of the power grid at time t; represents the target load of the power grid at time t; Indicates taking, and ΔL max The smaller value of ΔL max It is the preset maximum adjustment amount of the power grid load.

[0081] Further, the S6 includes:

[0082] Hierarchical division and definition: clarify the hierarchical structure of the charging system, which is divided into the level of individual charging stations, the regional level and the entire charging network level, and determine the responsibilities and control scope of each level;

[0083] Collaboratively control goal setting, set overall goals, and set specific sub-goals for each level;

[0084] Charging station-level collaborative control adjusts the charging power distribution of electric vehicles in the station according to the real-time power monitoring of the charging station and the charging demand determined by S2, giving priority to vehicles in urgent need of charging, ensuring that the total power does not exceed the limit, establishing an information sharing mechanism, coordinating charging resources, and avoiding local congestion;

[0085] Regional-level collaborative control aggregates data from charging stations in the region, formulates a charging resource allocation plan based on total charging demand and grid load, achieves charging power balance within the region, coordinates power flow between regions, ensures adjacent regions work together through intelligent scheduling algorithms, and achieves resource complementarity and optimal allocation;

[0086] The charging network is collaboratively controlled at the hierarchical level, integrating information uploaded by each area, using intelligent algorithms for global optimization, coordinating charging power between different areas, and ensuring the stability and efficiency of the entire network.

[0087] Furthermore, the calculation formula of the intelligent scheduling algorithm is:

[0088]

[0089] Where; k is the number of charging stations in the area, l is the number of electric vehicles in each charging station, P ij is the charging power of the i-th charging station in the j-th time period j , D is the total charging demand in the region, L is the upper limit of the grid load, SOC ij is the battery charge state of the i-th electric vehicle in the j-th time period, and the maximum charging power allowed by the battery is P max , α is the adjustment parameter; ε is a very small positive number.

[0090] Furthermore, the formula of the big data analysis algorithm is:

[0091]

[0092] P i ' j is the charging power of the optimized i-th charging station in the i-th time period; the required charging amount Q of the i-th electric vehicle in the j-th time period ij ; L is the upper limit of the grid load; F(Y) is the function of the big data analysis algorithm, which is used to predict the charging demand and grid load conditions, and adjust the charging power according to these prediction results; Y represents the relevant data vector input into the function.

[0093] In summary, due to the adoption of the above technical solution, the beneficial technical effects of the invention are:

[0094] Improve the overall efficiency of the charging system. Through multi-level optimization and control methods, it is possible to achieve coordinated control from a single charging station to the entire charging network, reasonably allocate charging resources, and improve the utilization rate of charging equipment, thereby significantly improving the overall efficiency of the charging system.

[0095] To achieve grid load balance, this method fully considers the load situation of the grid, coordinates the charging power between different areas through intelligent scheduling algorithms and big data analysis algorithms, avoids excessive load fluctuations on the grid caused by electric vehicle charging behavior, and effectively achieves grid load balance.

[0096] Reducing charging costs. Under the guidance of operating cost targets, by optimizing charging strategies and reasonably arranging charging time and power, we can reduce electricity procurement costs, charging facility construction and maintenance costs, etc., thereby reducing overall charging costs.

[0097] Improve user satisfaction. Considering the user satisfaction goal enables the charging strategy to take into account factors such as charging time, charging convenience and charging cost, providing users with more convenient and efficient charging services, thereby improving user satisfaction.

[0098] Enhance the stability and reliability of the charging system, implement real-time scheduling and control, and multi-level collaborative control, and make timely adjustments based on real-time data from electric vehicles and power grids to ensure the real-time, adaptability, and stability of the charging process, thus enhancing the reliability of the charging system.

[0099] In summary, the multi-level optimization and control method for large-scale electric vehicle charging and discharging rate of the present invention has many beneficial effects, and is of great significance for improving the performance of the charging system, promoting the development of the electric vehicle industry, and ensuring the stable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0100] Figure 1 A flowchart of the multi-level optimization and control method for charging and discharging rates of large-scale electric vehicles;

[0101] Figure 2 A flow chart of a method for determining charging and discharging power;

[0102] Figure 3 Flowchart of the method for constructing multi-level optimization objectives;

[0103] Figure 4 A flow chart of a method for dynamically adjusting charging power;

[0104] Figure 5 Flowchart of the multi-level collaborative control method. DETAILED DESCRIPTION

[0105] In order to make the purpose, technical solution and advantages of the invention more clear, the invention is further described in detail below in combination with the embodiments. It should be understood that the specific embodiments described here are only used to explain the invention and are not used to limit the invention.

[0106] like Figure 1 As shown, the multi-level optimization and control method for charging and discharging rate of large-scale electric vehicles includes the following steps:

[0107] S1. Modeling of electric vehicle charging and discharging loads: by analyzing the historical driving and charging data of electric vehicles, a daily mileage prediction model is established to predict the daily mileage of electric vehicles, and the charging demand of the daily mileage is calculated according to the daily mileage prediction model;

[0108] In S1, the method for establishing the daily mileage prediction model is: using a method combining time series analysis (ARIMA model) and machine learning (random forest model), and integrating the prediction results of the two models by weighted average method to obtain the final daily mileage prediction value:

[0109] D pr =ω1×D ar +ω2×D rf ;

[0110]

[0111] D rf = f[X];

[0112] Among them, ω1 and ω2 are weights; μ is a constant term, is the autoregressive coefficient, which reflects the influence of the past daily mileage on the current prediction value; p is the autoregressive order, which means that the daily mileage data at several past time points are considered; D t-1 is the daily mileage at time i in the past; ε t is the white noise error term; θ jis the moving average coefficient, q is the moving average order; f is the random forest model function; X is the input vector containing features such as holidays and weather.

[0113] In S1, the charging demand determined by the daily mileage = daily mileage × energy consumption per unit mileage; the energy consumption per unit mileage can be determined according to the specific model and historical data of the electric vehicle; for example, if the energy consumption of an electric vehicle is 0.2 kWh per kilometer and the predicted daily mileage is 100 kilometers, the power demand is 0.2×100=20 kWh.

[0114] S2. Determine the charging and discharging power capacity. Determine the charging demand of the daily mileage according to S1. Design an optimization model based on linear programming based on the determined charging demand to obtain the reasonable charging power of each electric vehicle in different time periods; achieve fair distribution of power capacity, ensure the power utilization efficiency of the charging station and the reasonable attenuation of the battery charging rate;

[0115] like Figure 2 As shown, the S2 includes:

[0116] S21. Goal setting, including fairness indicators, power efficiency indicators, and reasonable decay battery charging rates;

[0117] Among them, the goal of the fairness index is to minimize the difference between the charging powers of different electric vehicles to achieve fair distribution. The fairness index is defined as:

[0118]

[0119] F represents the fairness index, P sch,i represents the dispatchable power of the i-th electric vehicle; represents the average dispatchable power of all electric vehicles;

[0120] The goal of the power usage efficiency index is to maximize the power usage efficiency of the charging station. The power usage efficiency index is defined as:

[0121]

[0122] P i,j represents the charging power of the i-th electric vehicle in the j-th time period; Δt represents the length of the time period; P total,j represents the total power capacity of the charging station in the jth time period; m is the total number of time periods; n is the total number of electric vehicles;

[0123] The goal of a reasonable decay battery charging rate is to determine the power upper limit of different charging stages according to the battery's state of charge (SOC). The relationship between the battery's charging power and SOC is expressed by the following function:

[0124]

[0125] P ch (SOC i,j ) represents the maximum charging power of the i-th electric vehicle at a specific SOC in the j-th time period; P max is the maximum charging power allowed by the battery; SOC i,j is the battery state of charge of the i-th electric vehicle in the j-th time period; SOC max is the maximum state of charge of the battery; α is an adjustment parameter used to control the speed at which the charging power decreases as SOC increases;

[0126] S22. Design an optimization model based on linear programming:

[0127]

[0128] Among them, P sch represents the dispatchable power of a single electric vehicle; P max is the maximum charging power, which limits the upper limit of the charging power of each electric vehicle; P total is the total power of the charging station, reflecting the total power that the charging station can provide; n is the total number of electric vehicles; m is the total number of time periods; P i is the power allocated to the i-th electric vehicle; is an indicator function, which is 1 when the i-th electric vehicle is charging in the j-th time period, otherwise it is 0, indicating that the impact of the charging state of each electric vehicle in different time periods on the power allocation is taken into account;

[0129] S23. Determine the constraints:

[0130] (3) The sum of the charging power of all electric vehicles cannot exceed the total power capacity of the charging station, that is:

[0131]

[0132] P i,j represents the charging power of the i-th electric vehicle in the j-th time period; P total,j represents the total power of the charging station in the jth time period; n is the total number of electric vehicles; m is the total number of time periods;

[0133] (4) The charging power of each electric vehicle cannot exceed its maximum allowable value, that is:

[0134] 0≤P i，j ≤P i，max ;

[0135] P i,j represents the charging power of the i-th electric vehicle in the j-th time period; P i，maxis the maximum charging power of the i-th electric vehicle;

[0136] (3) Consider the power upper limit at different charging stages to avoid damage to battery life caused by overcharging. According to the charging characteristics of the battery, the charging power should meet the following requirements:

[0137] P i，j ≤P ch (SOC i,j );

[0138] P ch (SOC i,j ) represents the maximum charging power of the i-th electric vehicle at a specific SOC in the j-th time period; P i,j represents the charging power of the i-th electric vehicle in the j-th time period;

[0139] S24. Solve the model using a linear programming solver (such as CPLEX or Gurobi), input the objective function obtained in S21 and the constraints in S22 into the solver to obtain the reasonable charging power of each electric vehicle in different time periods.

[0140] like Figure 3 As shown, S3. Multi-level optimization target construction, based on the charging demand determined in S2, a multi-level optimization model is constructed;

[0141] The S3 includes:

[0142] S31. Determine optimization objectives, including operating cost objectives, user satisfaction objectives, and grid load balance objectives;

[0143] Among them, the cost function is expressed as:

[0144]

[0145] Where C(Q) represents the functional relationship between operating cost and charging demand; Q i represents the charging demand of the i-th electric vehicle; n represents the total number of electric vehicles; χ is the electricity procurement cost coefficient per unit charging demand;

[0146] The user satisfaction function is expressed as:

[0147]

[0148] Among them, U(Q) represents the functional relationship between user satisfaction and charging demand; ω represents the adjustment coefficient, P i represents the charging power of the i-th electric vehicle; Q i represents the charging demand of the i-th electric vehicle;

[0149] The grid load balancing objective is expressed as:

[0150]

[0151] Where B(Q) represents the functional relationship between grid load balance and charging demand; η is the adjustment coefficient; n represents the total number of electric vehicles; P i represents the charging power of the i-th electric vehicle; Q i represents the charging demand of the i-th electric vehicle; L max is the maximum value of grid load balance;

[0152] S32. Multi-level optimization objective function construction, integrating the optimization objectives determined in S31 into a model, constructing multi-level optimization objectives, and the multi-level optimization objective function is expressed as:

[0153] G(Q)=αC(Q)+βU(Q)+γB(Q);

[0154] G(Q) represents the multi-level optimization objective function; α, β, γ are weight coefficients used to adjust the importance of each objective; C(Q), U(Q), B(Q) represent the operating cost objective, user satisfaction objective, and grid load balance objective, respectively;

[0155] Objective function constraints:

[0156] Charging power constraints, Q total Is the charging demand online; Q s represents the charging demand of the sth electric vehicle; n is the total number of electric vehicles.

[0157] S4. Optimization of charging and discharging strategies. Under the guidance of the multi-objective optimization model constructed in S3, a heuristic algorithm is used to refine the specific charging and discharging strategies of each electric vehicle;

[0158] The S4 includes:

[0159] Set the initial parameters of the algorithm, such as population size, number of iterations, etc.

[0160] Representing the charging and discharging strategy of each electric vehicle as a coding form, such as a binary code or a real number code;

[0161] Randomly generate a set of initial charging and discharging strategies as the initial population of the algorithm;

[0162] Calculate the fitness value according to the multi-objective optimization function constructed by S3;

[0163] Among them, the fitness function is expressed as:

[0164]

[0165] In the formula, FP Fitness value; C max is the maximum value of operating cost; U max is the maximum value of user satisfaction; L max is the maximum value of the grid load balance; C(Q), U(Q), and L(Q) represent the operating cost target, user satisfaction target, and grid load balance target, respectively; ω C is the target weight of operating cost; U is the target weight of user satisfaction; ω B is the target weight for grid load balancing;

[0166] A set of initial charging and discharging strategies are randomly generated, each strategy is represented as a vector, which contains the charging power or discharging power of each electric vehicle in each time period; through the selection operation, excellent individuals are selected to enter the next generation according to the value of the fitness function; through the crossover operation, the genes of the selected individuals are exchanged to generate new individuals; through the mutation operation, some genes of the individuals are randomly changed to increase the diversity of the population;

[0167] Repeat the above steps until the termination condition is met, such as reaching the maximum number of iterations or finding a satisfactory solution.

[0168] like Figure 4 As shown, S5. Real-time dispatching and control, according to the charging and discharging strategy determined by S4, combined with the real-time data of electric vehicles and power grids, dynamically adjust the charging power;

[0169] The S5 includes:

[0170] S51. Collect relevant data of electric vehicles and power grids in real time through sensors, communication networks and other means, and transmit the real-time data to the control center; the real-time data includes parameters such as electric vehicle current, voltage, battery temperature, and frequency, voltage, power and other information of the power grid;

[0171] S52. According to the charging and discharging strategy determined in S4, understand the ideal charging power and time arrangement of each electric vehicle under different conditions;

[0172] S53. Analyze the data collected in real time to determine whether the actual demand of the current electric vehicle and the load of the power grid are consistent with the charging and discharging strategy of S4; if there is a difference, adjustments need to be made;

[0173] S54. Based on the analysis results of real-time data and combined with the charging and discharging strategy of S4, a real-time scheduling algorithm is used to dynamically adjust the charging power;

[0174] S55. Send the adjusted charging power instruction to the charging pile or charging equipment to ensure that they can charge according to the new power;

[0175] S56. During the charging process, the real-time data of the electric vehicle and the power grid are continuously monitored, and the charging power is continuously adjusted according to the actual situation to ensure the real-time, adaptability and stability of the charging process.

[0176] Furthermore, in S54, the real-time scheduling algorithm is expressed as:

[0177]

[0178] P c ' ,t represents the value of the adjusted charging power at time t; P c,t represents the current charging power value at time t; λ is the charging power adjustment coefficient; sign(x) is the sign function, when x>0, sign(x)=1; when x=0, sign(x)=0; when x<0, sign(x)=-1; represents the ideal charging power value at time t determined according to the charging and discharging strategy in S4; Indicates taking and ΔP max The smaller value of ΔP max is the preset maximum adjustment of charging power; μ is the grid load adjustment coefficient; represents the actual load of the power grid at time t; represents the target load of the power grid at time t; Indicates taking, and ΔL max The smaller value of ΔL max It is the preset maximum adjustment amount of the power grid load.

[0179] S6. Multi-level collaborative control, based on the dynamic adjustment results of S5, coordinates the charging power at different levels to achieve multi-level collaborative control, ensuring multi-level collaboration from a single charging station to the entire charging network.

[0180] like Figure 5 As shown, the S6 includes:

[0181] S61. Hierarchical division and definition: clarify the hierarchical structure of the charging system, divide the charging system into the single charging station level, the regional level (a region consisting of multiple charging stations) and the entire charging network level; determine the responsibilities and control scope of each level. For example, the single charging station level is responsible for the charging management of electric vehicles in the station; the regional level is responsible for coordinating the resource allocation of multiple charging stations in the region; and the entire charging network level is optimized and regulated from a global perspective;

[0182] S62. Collaborative control target setting, setting the overall target of multi-level collaborative control, such as ensuring the overall efficiency of the charging system is maximized, the grid load is balanced, and the charging cost is reduced; setting specific sub-targets for each level. For example, the sub-target at the level of a single charging station can be to improve the utilization rate and charging efficiency of the charging equipment in the station; the sub-target at the regional level can be to achieve balanced distribution of charging resources in the region and stable grid load; the sub-target at the level of the entire charging network can be to optimize the operating cost and reliability of the entire network.

[0183] S63. Collaborative control of charging stations. Based on the real-time power monitored by the charging stations and the charging demand determined by S2, a priority sorting algorithm is used to adjust the charging power distribution of electric vehicles in the station, giving priority to meeting the needs of electric vehicles that are in urgent need of charging, while ensuring that the total power of the charging station does not exceed its limit; an information sharing mechanism is established between charging stations, and charging resource information is exchanged with adjacent charging stations in real time through the communication network to coordinate charging resources and avoid local congestion;

[0184] S64. Regional-level collaborative control, summarizing the power and load data uploaded by each charging station in the region, formulating a reasonable charging resource allocation plan based on the total charging demand and grid load in the region, and achieving a balance of charging power in the region; coordinating power flows between regions, and ensuring that adjacent regions can work together through intelligent scheduling algorithms to achieve complementary and optimal allocation of resources;

[0185] S65. The charging network is coordinated at the hierarchical level, integrating the information uploaded by each area and using big data analysis algorithms to perform global optimization from the perspective of the entire charging network. Through big data analysis algorithms, the charging power between different areas is coordinated to ensure the stability and efficiency of the entire network.

[0186] In step 64, the calculation formula of the intelligent scheduling algorithm is:

[0187]

[0188] Formula meaning: On the basis of considering the total charging demand and grid load in the region, the battery state of charge of the electric vehicle and the maximum charging power allowed by the battery are further considered. The charging power is proportional to the total charging demand and the remaining capacity of the grid load, and decreases with the increase of the battery state of charge to avoid overcharging. In addition, the maximum charging power allowed by the battery is also considered to ensure that the charging power is within a reasonable range.

[0189] Where k is the number of charging stations in the region, l is the number of electric vehicles in each charging station, and the charging power of the i-th charging station in the j-th time period is P ij, D is the total charging demand in the region, L is the upper limit of the grid load, and the battery state of charge of the i-th electric vehicle in the j-th time period is SOC ij , the maximum charging power allowed by the battery is P max , α is the adjustment parameter; ε is a very small positive number.

[0190] The formula of the big data analysis algorithm in S65 is:

[0191]

[0192] P i ' j is the charging power of the i-th charging station in the i-th time period after optimization; Q ij is the required charging amount of the i-th electric vehicle in the j-th time period; L is the upper limit of the grid load; F(Y) is the function of the big data analysis algorithm; it takes into account various factors, such as the historical charging data of electric vehicles, the historical data of grid load, weather conditions, time, etc., to predict the charging demand and grid load conditions, and adjust the charging power according to these prediction results; Y represents the relevant data vector input into the function.

[0193] The above description is a preferred embodiment of the invention and is not intended to limit the invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the invention should be included in the protection scope of the invention.

Claims

1. A multi-level optimization control method for charging and discharging rates of large-scale electric vehicles, characterized in that: include: By analyzing the historical driving and charging data of electric vehicles, a daily mileage prediction model is established, and the charging demand of daily mileage is calculated based on the daily mileage prediction model; Based on the determined charging demand, an optimization model based on linear programming is designed to obtain the reasonable charging power of each electric vehicle in different time periods; Based on the determined charging demand, a multi-level optimization model is constructed; Under the guidance of the constructed multi-objective optimization model, a heuristic algorithm is used to refine the specific charging and discharging strategy of each electric vehicle; according to the determined charging and discharging strategy, the charging power is dynamically adjusted in combination with the real-time data of the electric vehicle and the power grid; based on the results of the dynamic adjustment of the charging power, the charging power at different levels is coordinated to achieve multi-level collaborative control.

2. The multi-level optimization control method for charging and discharging rate of large-scale electric vehicles according to claim 1 is characterized in that: The method for establishing the daily mileage prediction model includes: By combining time series analysis with machine learning, the prediction results of the two models are merged through the weighted average method to obtain the final daily mileage prediction value: D pr =ω1×D ar +ω2×D rf ; D rf =f[X]; Among them, ω1 and ω2 are weights; μ is a constant term, is the autoregressive coefficient, which reflects the influence of the past daily mileage on the current predicted value; p is the autoregressive order, which means that the daily mileage data at several points in the past are considered; D t-1 is the daily mileage at time i in the past; ε t is the white noise error term; θ j is the moving average coefficient, q is the moving average order; f is the random forest model function; X is the input vector containing features such as holidays and weather.

3. The multi-level optimization control method for charging and discharging rate of large-scale electric vehicles according to claim 1 is characterized in that: The charging demand determined by the daily driving mileage=daily driving mileage×energy consumption per unit mileage.

4. The multi-level optimization control method for charging and discharging rate of large-scale electric vehicles according to claim 1 is characterized in that: Designing an optimization model based on linear programming involves: S21. Goal setting, including fairness indicators, power efficiency indicators, and reasonable decay battery charging rates; Among them, the goal of the fairness index is to minimize the difference between the charging powers of different electric vehicles to achieve fair distribution. The fairness index is defined as: F represents the fairness index, P sch,i represents the dispatchable power of the i-th electric vehicle; represents the average dispatchable power of all electric vehicles; The goal of the power usage efficiency index is to maximize the power usage efficiency of the charging station. The power usage efficiency index is defined as: P i,j represents the charging power of the i-th electric vehicle in the j-th time period; Δt represents the length of the time period; P total,j represents the total power of the charging station in the jth time period; m is the total number of time periods; n is the total number of electric vehicles; The goal of a reasonable decay battery charging rate is to determine the power upper limit of different charging stages according to the battery's state of charge (SOC). The relationship between the battery's charging power and SOC is expressed by the following function: P ch (SOC i,j ) represents the maximum charging power of the i-th electric vehicle at a specific SOC in the j-th time period; P max is the maximum charging power allowed by the battery; SOC i,j is the battery state of charge of the i-th electric vehicle in the j-th time period; SOC max is the maximum state of charge of the battery; α is an adjustment parameter used to control the speed at which the charging power decreases as SOC increases; S22. Design an optimization model based on linear programming: Among them, P sch represents the dispatchable power of a single electric vehicle; P max is the maximum charging power, which limits the upper limit of the charging power of each electric vehicle; P total is the total power of the charging station, reflecting the total power that the charging station can provide; n is the total number of electric vehicles; m is the total number of time periods; P i is the power allocated to the i-th electric vehicle; is an indicator function, which is 1 when the i-th electric vehicle is charging in the j-th time period, otherwise it is 0, indicating that the impact of the charging state of each electric vehicle in different time periods on the power allocation is taken into account; S23. Determine the constraints: (1) The sum of the charging power of all electric vehicles cannot exceed the total power capacity of the charging station, that is: P i,j represents the charging power of the i-th electric vehicle in the j-th time period; P total,j represents the total power capacity of the charging station in the jth time period; n is the total number of electric vehicles; m is the total number of time periods; (2) The charging power of each electric vehicle cannot exceed its maximum allowable value, that is: 0≤P i，j ≤P i，max ; P i,j represents the charging power of the i-th electric vehicle in the j-th time period; P i，max is the maximum charging power of the i-th electric vehicle; (3) Consider the power upper limit at different charging stages to avoid damage to battery life caused by overcharging. According to the charging characteristics of the battery, the charging power should meet the following requirements: P i，j ≤P ch (SOC i,j ); P ch (SOC i,j ) represents the maximum charging power of the i-th electric vehicle at a specific SOC in the j-th time period; P i,j represents the charging power of the i-th electric vehicle in the j-th time period; S24. Solve the model using a linear programming solver (such as CPLEX or Gurobi), input the objective function obtained in S21 and the constraints in S22 into the solver to obtain the reasonable charging power of each electric vehicle in different time periods.

5. The multi-level optimization control method for charging and discharging rate of large-scale electric vehicles according to claim 1 is characterized in that: Methods for building multi-level optimization models: S31. Determine optimization objectives, including operating cost objectives, user satisfaction objectives, and grid load balance objectives; Among them, the cost function is expressed as: Where C(Q) represents the functional relationship between operating cost and charging demand; Q i represents the charging demand of the i-th electric vehicle; n represents the total number of electric vehicles; χ is the electricity procurement cost coefficient per unit charging demand; The user satisfaction function is expressed as: Among them, U(Q) represents the functional relationship between user satisfaction and charging demand; ω represents the adjustment coefficient, P i represents the charging power of the i-th electric vehicle; Q i represents the charging demand of the i-th electric vehicle; The grid load balancing objective is expressed as: Where B(Q) represents the functional relationship between grid load balance and charging demand; η is the adjustment coefficient; n represents the total number of electric vehicles; P i represents the charging power of the i-th electric vehicle; Q i represents the charging demand of the i-th electric vehicle; L max is the maximum value of grid load balance; S32. Multi-level optimization objective function construction, integrating the optimization objectives determined in S31 into a model, constructing multi-level optimization objectives, and the multi-level optimization objective function is expressed as: G(Q)=αC(Q)+βU(Q)+γB(Q); G(Q) represents the multi-level optimization objective function; α, β, γ are weight coefficients used to adjust the importance of each objective; C(Q), U(Q), B(Q) represent the operating cost objective, user satisfaction objective, and grid load balance objective, respectively; Objective function constraints: Charging power constraints, Q total Is the charging demand online; Q i represents the charging demand of the i-th electric vehicle; n is the total number of electric vehicles.

6. The multi-level optimization control method for charging and discharging rate of large-scale electric vehicles according to claim 1 is characterized in that: The specific charging and discharging strategies for each electric vehicle are refined using heuristic algorithms, including: Set the initial parameters of the algorithm, such as population size and number of iterations; Representing the charging and discharging strategy of each electric vehicle as a coding form, such as a binary code or a real number code; Randomly generate a set of initial charging and discharging strategies as the initial population of the algorithm; Calculate the fitness value according to the multi-objective optimization function constructed by S3; Among them, the fitness function is expressed as: In the formula, F P Fitness value; C max is the maximum value of operating cost; U max is the maximum value of user satisfaction; L max is the maximum value of the grid load balance; C(Q), U(Q), and L(Q) represent the operating cost target, user satisfaction target, and grid load balance target, respectively; ω C is the target weight of operating cost; U is the target weight of user satisfaction; ω B is the target weight for grid load balancing; A set of initial charging and discharging strategies are randomly generated, each strategy is represented as a vector, which contains the charging power or discharging power of each electric vehicle in each time period; through the selection operation, excellent individuals are selected to enter the next generation according to the value of the fitness function; through the crossover operation, the genes of the selected individuals are exchanged to generate new individuals; through the mutation operation, some genes of the individuals are randomly changed to increase the diversity of the population; The above operations are repeated until the termination condition is met, such as reaching the maximum number of iterations or finding a satisfactory solution.

7. The multi-level optimization control method for charging and discharging rate of large-scale electric vehicles according to claim 1 is characterized in that: The dynamically adjusting charging power comprises: Through sensors, communication networks and other means, relevant data of electric vehicles and power grids are collected in real time and transmitted to the control center; real-time data includes electric vehicle current, voltage, battery temperature parameters, and power grid frequency, voltage and power information; According to the charging and discharging strategies determined by S4, understand the ideal charging power and time schedule for each electric vehicle under different circumstances; Analyze the data collected in real time to determine whether the actual demand of electric vehicles and the load of the power grid are consistent with the charging and discharging strategy of S4; if there is a difference, adjustments need to be made; According to the analysis results of real-time data, combined with S4's charging and discharging strategy, a real-time scheduling algorithm is used to dynamically adjust the charging power; Send the adjusted charging power command to the charging pile or charging equipment to ensure that they can charge according to the new power; During the charging process, the real-time data of electric vehicles and power grids are continuously monitored, and the charging power is continuously adjusted according to the actual situation to ensure the real-time, adaptability and stability of the charging process; The real-time scheduling algorithm is expressed as: P c ' ,t represents the value of the adjusted charging power at time t; P c,t represents the current charging power value at time t; λ is the charging power adjustment coefficient; sign(x) is the sign function, when x>0, sign(x)=1; when x=0, sign(x)=0; when x<0, sign(x)=-1; represents the ideal charging power value at time t determined according to the charging and discharging strategy in S4; Indicates taking and ΔP max The smaller value of ΔP max is the preset maximum adjustment of charging power; μ is the grid load adjustment coefficient; represents the actual load of the power grid at time t; represents the target load of the power grid at time t; Indicates taking, and ΔL max The smaller value of ΔL max It is the preset maximum adjustment amount of the power grid load.

8. The multi-level optimization and control method for charging and discharging rate of large-scale electric vehicles according to claim 1 is characterized in that: Methods for achieving multi-level collaborative control: Hierarchical division and definition: clarify the hierarchical structure of the charging system, which is divided into the level of individual charging stations, the regional level and the entire charging network level, and determine the responsibilities and control scope of each level; Collaboratively control goal setting, set overall goals, and set specific sub-goals for each level; Charging station-level collaborative control adjusts the charging power distribution of electric vehicles in the station according to the real-time power monitoring of the charging station and the charging demand determined by S2, giving priority to vehicles in urgent need of charging, ensuring that the total power does not exceed the limit, establishing an information sharing mechanism, coordinating charging resources, and avoiding local congestion; Regional-level collaborative control aggregates data from charging stations in the region, formulates a charging resource allocation plan based on total charging demand and grid load, achieves charging power balance within the region, coordinates power flow between regions, ensures adjacent regions work together through intelligent scheduling algorithms, and achieves resource complementarity and optimal allocation; The charging network is collaboratively controlled at the hierarchical level, integrating information uploaded by each area, using intelligent algorithms for global optimization, coordinating charging power between different areas, and ensuring the stability and efficiency of the entire network.

9. The multi-level optimization control method for charging and discharging rate of large-scale electric vehicles according to claim 8 is characterized in that: The calculation formula of the intelligent scheduling algorithm is: Where k is the number of charging stations in the region, l is the number of electric vehicles in each charging station, and the charging power of the i-th charging station in the j-th time period is P ij , D is the total charging demand in the region, L is the upper limit of the grid load, and the battery state of charge of the i-th electric vehicle in the j-th time period is SOC ij , the maximum charging power allowed by the battery is P max , α is the adjustment parameter; ε is a very small positive number.

10. The multi-level optimization control method for charging and discharging rate of large-scale electric vehicles according to claim 8, characterized in that: The formula of the intelligent algorithm is: P i ' j is the charging power of the i-th charging station in the j-th time period after optimization; Q ij is the required charging amount of the i-th electric vehicle in the j-th time period; L is the upper limit of the grid load; F(Y) is the function of the big data analysis algorithm; Y represents the relevant data vector input into the function.

Images