Electric vehicle-energy storage collaborative charging and discharging method based on multi-objective optimization

Abstract

The method comprises the following steps: building a community micro-grid system; using a peak-valley electricity price strategy to control the charging and discharging behavior of the electric vehicle, so that the energy storage and the electric vehicle are coordinated to be connected to the grid, and the energy storage and the electric vehicle are discharged during the peak time and charged during the valley time; after the energy storage and the electric vehicle are coordinated to be connected to the grid, a target optimization model is established, which considers that the load peak-valley difference is minimum, the grid fluctuation mean square error is minimum, and the user's expenditure is minimum; the target optimization model is solved according to an improved non-dominated sorting genetic algorithm, and a Pareto non-inferior solution set satisfying the requirements is obtained; in the Pareto non-inferior solution set, an optimal solution is selected from the optimal solution set according to a fuzzy membership function, and is used as a planned output value of the energy storage and the electric vehicle. The method realizes that the energy storage and the electric vehicle participate in V2G coordinated optimization scheduling, can effectively balance the peak-valley load of the system, and can reduce the user's expenditure.

Description

Technical Field

[0001] This invention relates to the field of microgrid optimization scheduling, and specifically to a method for coordinated charging and discharging of electric vehicles and energy storage based on multi-objective optimization. Background Technology

[0002] Electric vehicles, as a new type of bidirectional electrical load, can function as both mobile energy storage devices and power-consuming loads. Given that both energy storage systems and electric vehicle V2G technology have charging and discharging potential, they can be flexibly scheduled to achieve effects such as smoothing load curves and reducing peak-to-valley differences.

[0003] In community microgrids, the orderly charging and discharging of energy storage and electric vehicles can stabilize the load balance in the power grid and provide more flexible energy services for both users and the grid. Current research on this issue mainly focuses on the following two aspects:

[0004] 1) Energy storage capacity optimization configuration issues:

[0005] Currently, researchers primarily address this problem through optimization algorithms and models. These include linear programming models, multi-objective optimization models, genetic algorithms, ant colony optimization, and particle swarm optimization. By comparing the advantages and disadvantages of various algorithms, researchers can derive the optimal energy storage capacity configuration strategy, thereby maximizing grid benefits and minimizing costs.

[0006] 2) The problem of orderly charging and discharging of electric vehicles:

[0007] To address the challenges posed by disorderly charging of electric vehicles, researchers are currently exploring orderly charging solutions based on smart charging technologies. For example, developing charging scheduling algorithms based on vehicle-to-everything (V2X) communication technology enables electric vehicles to charge and discharge in an orderly manner during peak and off-peak periods, ensuring the stability and security of the power grid. Furthermore, based on artificial intelligence and machine learning technologies, researchers can also develop more intelligent charging scheduling strategies to provide users with better energy services.

[0008] As research on microgrid energy management deepens, establishing optimal energy management strategies to achieve optimal energy allocation remains a pressing issue. Furthermore, the uncertainty of load demand, the randomness of distributed energy output, and the disorderly charging behavior of electric vehicles can also cause serious power quality problems in community microgrids. Therefore, conducting research on the coordinated optimization scheduling of source-grid-load-storage community microgrids containing electric vehicles has significant practical implications. Summary of the Invention

[0009] To address the shortcomings of existing technologies, this invention proposes a multi-objective optimization-based electric vehicle-energy storage collaborative charging and discharging method. In a photovoltaic-load-storage microgrid system, this method establishes a day-ahead optimization scheduling model with three objectives: user expenditure, peak-valley power difference, and mean square error. An improved non-dominated sorting genetic algorithm is used to solve for the planned output values ​​of energy storage and electric vehicles. By leveraging the charging and discharging characteristics of electric vehicles and energy storage to interact bidirectionally with the grid, their charging and discharging potential is utilized for discharging during peak load periods and charging energy storage during off-peak periods, thereby smoothing the load curve and reducing the peak-valley difference. This method enables energy storage and electric vehicles to collaboratively participate in optimized scheduling, effectively balancing the system's peak and valley loads while reducing user expenditure.

[0010] The technical solution adopted in this invention is as follows:

[0011] The electric vehicle-energy storage coordinated charging and discharging method based on multi-objective optimization includes the following steps:

[0012] Step 1: Construct a community microgrid system that includes photovoltaic power generation systems, energy storage systems, electric vehicles, and the electricity load of community users;

[0013] Step 2: Use peak-valley electricity pricing strategy to control the charging and discharging behavior of electric vehicles, so that energy storage and electric vehicles can be connected to the grid in a coordinated manner, discharging during peak hours and charging during off-peak hours;

[0014] Step 3: After energy storage and electric vehicles are connected to the grid in a coordinated manner, establish an objective optimization model that considers minimizing the load peak-valley difference, minimizing the mean square error of grid connection fluctuation, and minimizing user expenditure costs;

[0015] Step 4: Solve the objective optimization model established in Step 3 using the improved Nondominated Sorting Genetic Algorithm (NSGA-II) to obtain the set of Pareto nondominated solutions that meet the requirements;

[0016] Step 5: From the Pareto non-dominated solution set, select the optimal solution from the optimal solution set according to the fuzzy membership function, and use it as the planned output value for energy storage and electric vehicles.

[0017] In step one:

[0018] A: Photovoltaic output is mainly related to light intensity. This invention uses a beta distribution model to fit the distribution function of light intensity, and its probability density is:

[0019]

[0020] In the formula, Γ represents the gamma function; r represents the actual measured light intensity; r maxα represents the maximum light intensity; β and α are two shape parameters of the beta distribution.

[0021] Its photovoltaic output power expression is:

[0022] P PV =η×r×A

[0023] In the formula, P PV η represents the total output power of the photovoltaic array; A represents the conversion efficiency of the photovoltaic array; and η represents the total area of ​​the photovoltaic array.

[0024] b: Energy storage has flexible charging and discharging characteristics, which can be used to achieve peak power consumption and smooth load curves.

[0025] The relationship between the energy storage system's capacity and its charging / discharging power is expressed as follows:

[0026] C ES (t)=C ES (t-1)+P s (t)△t

[0027] In the formula, C ES (t-1) and C ES (t) represents the amount of electricity stored in the energy storage system during time periods t-1 and t, respectively; P s (t) represents the power that the energy storage system can output at time t; Δt is the time interval between time t-1 and time t.

[0028] The charging and discharging process of an energy storage system is as follows:

[0029] (1) Discharge process:

[0030] The discharge power of the energy storage system in hour t on day j is:

[0031]

[0032] In the formula, P ed C is the rated discharge power of the energy storage system. ESmax C represents the maximum capacity of the energy storage system. ES (t-1,j) represents the energy storage capacity of the energy storage system during the (t-1)th time period on day j.

[0033] C ES (t,j) is affected by its own energy storage capacity, and its expression is:

[0034]

[0035] In the formula, C ES (t,j) represents the energy storage capacity of the energy storage system at time t on day j; C ESminThis represents the minimum capacity of the energy storage system.

[0036] C ESV (t,j) represents the ideal capacity of the energy storage system, expressed as:

[0037]

[0038] In the formula, η ed The discharge efficiency of the energy storage system;

[0039] (2) Charging process:

[0040] The charging power of the energy storage system in hour t on day j is:

[0041]

[0042] In the formula, P ec Rated charging power for the energy storage system.

[0043] C ES (t,j)=C ES (t-1,j)+△tP es (t,j)η es

[0044] In the formula, η es The charging efficiency of the energy storage system; P es (t,j) represents the charging power of the energy storage system at time t on day j.

[0045] In step two:

[0046] The random charging and discharging behavior of a large number of electric vehicles connected to the microgrid may further increase the peak-valley difference of the load curve. This invention adopts a peak-valley electricity pricing strategy to guide the charging and discharging behavior of electric vehicle owners. By taking advantage of the fact that electric vehicles can be both loads and mobile energy storage, load fluctuations can be smoothed out and the charging costs for owners can be reduced.

[0047] Due to the behavioral characteristics of electric vehicle owners, such as the randomness of their departure and return times, the discharge process is divided into daytime peak periods (6:00-16:00) and evening peak periods (after 16:00), and two discharge strategies are formulated:

[0048] (1). The actual discharge power of all electric vehicles participating in the time-of-use pricing response during the daytime peak period t hour on day j is:

[0049]

[0050] In the formula, N d(t) represents the number of electric vehicles participating in discharge during the daytime peak period; i represents the i-th electric vehicle responding to the time-of-use electricity price.

[0051] Let be the daytime peak discharge power of the i-th electric vehicle, expressed as:

[0052]

[0053] In the formula, η evd SOC represents discharge efficiency; SOC0 represents the electric vehicle's state of full charge; SOC min This represents the minimum state of charge for discharge; C EV For electric vehicle capacity.

[0054]

[0055] In the formula, p B ξ represents the proportion of electric vehicles in an idle state; ξ is the user participation coefficient, representing the proportion of electric vehicle owners who respond to the time-of-use pricing strategy; N EV m is the total number of electric vehicles. d This refers to the number of peak daytime hours.

[0056] (2). The actual discharge power of all electric vehicles participating in the time-of-use pricing response during the peak period t hour on day j is:

[0057]

[0058] in:

[0059]

[0060]

[0061]

[0062] In the formula, m n p represents the number of evening peak hours; A For electric vehicles that are in use and responding to time-of-use pricing; f s (t) is the probability function of the electric vehicle's return time; N n (t) represents the number of electric vehicles participating in discharge during the evening peak period.

[0063] Electric vehicle charging strategies:

[0064] On day j, during the Pinggu period t hours, the electric vehicle is charged, and the amount of charge is:

[0065]

[0066] In the formula, C EVc(t,j) represents the charging load during the t-hour period in the valley. Due to the time-of-use pricing, large-scale electric vehicles exhibit similar charging behavior characteristics, which can be simulated using the Monte Carlo method.

[0067] In step two:

[0068] By setting parameters λ1 and λ2 to adjust the ratio of energy storage system and electric vehicle grid connection during peak and valley periods respectively, the model for their coordinated grid connection output is as follows:

[0069] P d (t,j)=λ1P esd (t,j)+(1-λ1)P evd (t,j) t∈peak period

[0070] P c (t,j)=λ2P esc (t,j)+(1-λ2)P evc (t,j) t∈Pinggu period

[0071] In the formula, P d (t,j) represents the discharge power of the energy storage and electric vehicle during time period t on day j; P c (t,j) represents the energy storage and electric vehicle charging power during time period t on day j; P esd (t,j) represents the discharge power of the energy storage system during time period t on day j; P evd (t,j) represents the discharge power of the electric vehicle participating in the time-of-use electricity pricing response during time period t on day j; P esc (t,j) represents the charging power of the energy storage system during time period t on day j; P evc (t,j) represents the charging power of the electric vehicle during time period t on day j.

[0072] Here, it is assumed that the total power provided by all electric vehicles participating in the time-of-use pricing response during peak hours is P. EVD The total power that all energy storage devices can provide is P. ESD Then we can obtain:

[0073]

[0074] Similarly, the values ​​of λ2 during the flat valley period are as follows:

[0075] Assume that the total power required for charging of all electric vehicles participating in time-of-use pricing during off-peak hours is P. EVC The total power required by all energy storage devices is P. ESC Then we can obtain:

[0076]

[0077] After energy storage and electric vehicles are connected to the grid in a coordinated manner, the time-series load P in the t-th time period of day j is... LA (t,j) is:

[0078]

[0079] In the formula, P L (t,j), P PV (t,j) represent the original load power and photovoltaic output power at time t on day j, respectively.

[0080] In step three, with the goal of achieving the best peak-shaving and valley-filling effect on the load curve after the coordinated grid connection of energy storage and electric vehicles, the following optimization objective function f1, which considers minimizing the load peak-valley difference, and the optimization objective function f2, which considers minimizing the root mean square error of grid connection fluctuation, are established:

[0081] f1=min(P LAmax (j)-P LAmin (j))

[0082]

[0083] In the formula, P LAmax (j), P LAmin (j) represent the maximum and minimum time-series load power on day j, respectively; t N This represents the number of time periods in a day.

[0084] Responding to time-of-use pricing for orderly charging and discharging will, to some extent, alter users' original behavioral characteristics. From an economic perspective, considering the implementation of an orderly charging and discharging strategy, an optimization objective function f3 is established to minimize user expenses.

[0085]

[0086] In the formula, δ t Electricity price for charging at different times; χ t =0 or 1, representing non-charging and charging states respectively.

[0087] The constraints on the above objective function are:

[0088] ① Photovoltaic output constraints:

[0089] P PVmin ≤P PV (t,j)≤P PVmin ;

[0090] P PVmin P PVmin These represent the upper and lower limits of photovoltaic power output, respectively; P PV (t,j) represents the output power of the photovoltaic system during the t-th time period on day j.

[0091] ② Capacity constraints of energy storage:

[0092] C ESmin ≤C ES (t)≤C ESmax ;

[0093] ③ Safety constraints for electric vehicle batteries:

[0094] SOC min ≤SOC i (t,j)≤SOC max ;

[0095] SOC min Indicates the lower limit of the battery's state of charge; SOC i (t,j) represents the state of charge (SOC) of the i-th electric vehicle during time period t on day j; max This indicates the upper limit of the battery's state of charge.

[0096] ④ Power balance constraints:

[0097] P grid (t,j)=P LA (t,j);

[0098] In the formula, P grid (t,j) represents the power supplied by the power grid during time period t on day j; P LA (t,j) represents the time-series load of time period t on day j. Step four includes the following steps:

[0099] Step a: Based on the problem of coordinated charging and discharging control of community microgrid energy storage and electric vehicles, N individuals are randomly generated. The basic algorithm parameters are input, initial crossover and mutation probabilities are set, and each individual is encoded. A Cubic chaotic mapping strategy is adopted, resulting in a more uniform initial population distribution and smaller differences between individuals. The Cubic chaotic mapping function is as follows:

[0100]

[0101] In the formula, x k Let x be the position of the k-th individual; a and b are chaos influencing factors; the Cubic mapping range depends on the values ​​of a and b, and when b∈(2.3,3), its mapping state is a chaotic sequence; x k+1 This represents the position of the (k+1)th individual mapped by the chaotic mapping of the kth individual.

[0102] Step b: Calculate the fitness value and constraint violation value of each individual in the population;

[0103] Step c: Rank individuals based on their merits using an improved non-dominated ranking strategy and their crowding distance;

[0104] Step d: Select dominant individuals as parent populations;

[0105] Step e: Perform selection, crossover, and mutation operations on the parent generation to generate the offspring population;

[0106] Step f: Merge the elite individuals selected by the elite strategy in the parent population with the offspring population generated in step e;

[0107] Step g: Recalculate the individual fitness values ​​and constraint violation values ​​of the newly generated population;

[0108] Step h: Determine if the maximum number of iterations has been reached. If yes, output the Pareto non-dominated solution set; otherwise, repeat step c.

[0109] Step c includes an improved non-dominated sorting strategy:

[0110] A new individual performance evaluation index—dominance coupling degree δ—is introduced: The smaller the dominance coupling degree of an individual in the non-dominated set, the better its optimization performance in most sub-objectives of the multi-objective optimization problem. This indicates that the individual's performance is superior and more likely to be passed on to its offspring. Its expression is:

[0111]

[0112] In the formula, individual A k It belongs to the non-dominated set; S is the number of sub-objects; and f represents the maximum and minimum fitness values ​​of the s-th sub-objective; s (A k ) represents individual A k The fitness value.

[0113] At this time, individual A k There is a non-dominant hierarchy Dominant Coupling Crowded distance Three attributes;

[0114] For any two individuals A k B k ,like or or Then individual A k Its performance is superior.

[0115] Step e includes introducing an adaptive crossover and mutation operator: the crossover and mutation probabilities of the traditional NSGA-II algorithm are all constant values, which may destroy excellent individuals and cause the population to "degenerate".

[0116]

[0117] In the formula, ρ c ρ m , respectively, represent the crossover probability and mutation probability; G and g represent the total number of evolutions and the current number of evolutions, respectively.

[0118] Step five is specifically as follows:

[0119] First, traverse the Pareto solution set and calculate the membership degree of the s-th objective function in the k-th solution:

[0120]

[0121] f smin f smax Let represent the minimum and maximum fitness values ​​of the s-th objective function, respectively.

[0122] Decision-makers can adjust the preference coefficients, i.e., the weights of each objective, based on the load situation. A weighted method is used to obtain the membership weights of the optimal solution problem. The Pareto solution corresponding to the maximum value obtained is the optimal solution to the problem. The calculation formula is as follows:

[0123]

[0124] In the formula, ρ s The preference coefficients of the s-th sub-objective function; N is the population size; is the membership function value; S represents the total number of objective functions; s represents the s-th objective function.

[0125] This invention provides a multi-objective optimization-based method for coordinated charging and discharging of electric vehicles and energy storage, with the following technical advantages:

[0126] 1) In step one of this invention, the research object is a community microgrid system that includes a photovoltaic power generation system, an energy storage system, electric vehicles, and the electricity load of community users. This community microgrid system, which integrates a photovoltaic power generation system, an energy storage system, electric vehicles, and the electricity load of community users, has many technical benefits, including improved energy utilization efficiency, optimized energy balance, improved power supply reliability, reduced environmental pollution, and increased economic benefits. This integrated energy management system is expected to play an important role in the future urban energy transition.

[0127] 2) In step two of this invention, a peak-valley electricity pricing strategy is adopted to guide the charging and discharging behavior of electric vehicle owners, enabling energy storage and electric vehicles to discharge during peak hours and charge during off-peak hours in a coordinated manner. By using this strategy to guide the charging and discharging behavior of electric vehicle owners and setting parameters λ1 and λ2 to adjust the grid connection ratio of energy storage systems and electric vehicles during peak and off-peak periods, coordinated operation of energy storage and electric vehicles is achieved. This optimizes energy utilization and supply-demand balance, improves the stability of the power system, enhances economic efficiency, and promotes sustainable energy. This strategy contributes to building a more intelligent, efficient, and environmentally friendly energy management system.

[0128] 3) In step three of this invention, after the coordinated grid connection of energy storage and electric vehicles, a three-objective optimization model is established, considering minimizing the load peak-valley difference, minimizing the mean square error of grid connection fluctuations, and minimizing user expenses. To maximize energy utilization and user economic benefits, a three-objective optimization model comprehensively considers the load peak-valley difference, the mean square error of grid connection fluctuations, and user expenses. The goal of this model is to achieve the above three optimization objectives by rationally scheduling the charging and discharging behavior of energy storage and electric vehicles while meeting electricity demand. Solving this model involves information from power system load forecasting, energy storage and electric vehicle state management, and the application of multi-objective optimization algorithms. Through the optimization results of this model, coordinated grid connection of energy storage and electric vehicles can be achieved, maximizing energy utilization and user economic benefits.

[0129] 4) In step four of this invention, the three-objective optimization problem is solved using the improved non-dominated sorting genetic algorithm to obtain a set of Pareto non-dominated solutions that satisfy the three objectives. This invention introduces a new individual performance evaluation index—dominance coupling degree δ—and uses sorting to find individuals with better performance for genetic mutation operations. Therefore, each solution in this set achieves relatively good performance on all three optimization objectives, providing decision-makers with diverse choices.

[0130] 5) In step five of this invention, from the set of three-objective Pareto non-dominated solutions, the optimal solution is selected from the set of optimal solutions as the planned output value for energy storage and electric vehicles based on the fuzzy membership function. In the set of three-objective Pareto non-dominated solutions, if the decision-maker wishes to select an optimal solution as the planned output value for energy storage and electric vehicles, the fuzzy membership function can be used for decision-making. The fuzzy membership function can be used to quantify the superiority of each solution in each objective, thereby assisting in making the final decision. Attached Figure Description

[0131] Figure 1 This is an overall block diagram of the method of the present invention;

[0132] Figure 2This is a flowchart of the improved non-dominated sorting genetic algorithm.

[0133] Figure 3 It is a graph showing the community load demand, corresponding time-of-use electricity prices, and photovoltaic output.

[0134] Figure 4 This is the Pareto front plot for a three-objective function.

[0135] Figure 5 It is a power allocation diagram under multi-objective optimization.

[0136] Figure 6 It is a diagram showing the charging and discharging power and state of charge of an energy storage system under multi-objective optimization.

[0137] Figure 7 This is a comparison chart of the electricity purchase situation of community microgrids from the power grid in four scenarios. Detailed Implementation

[0138] This invention proposes a multi-objective optimization-based electric vehicle-energy storage coordinated charging and discharging strategy, such as... Figure 1 As shown, the method includes:

[0139] Step 101: The research object of this invention is a community microgrid system that includes a photovoltaic power generation system, an energy storage system, electric vehicles, and the electricity load of community users;

[0140] Step 102: Use peak-valley electricity pricing strategy to guide the charging and discharging behavior of electric vehicle owners, so that energy storage and electric vehicles can discharge during peak hours and charge during off-peak hours in a coordinated manner and be connected to the grid.

[0141] Step 103: After energy storage and electric vehicles are connected to the grid in a coordinated manner, establish a three-objective optimization model that considers minimizing the load peak-valley difference, minimizing the mean square error of grid connection fluctuation, and minimizing user expenditure costs.

[0142] Step 104: Solve the three-objective optimization problem using the improved non-dominated sorting genetic algorithm to obtain the set of Pareto non-dominated solutions that satisfy the three objectives;

[0143] Step 105: In the set of non-dominated Pareto solutions for the three objectives, select the optimal solution from the set of optimal solutions based on the fuzzy membership function as the planned output value for energy storage and electric vehicles.

[0144] Example:

[0145] In a specific embodiment of this invention, the research object is a community microgrid system comprising a photovoltaic power generation system, an energy storage system, electric vehicles, and the electricity load of community users. This microgrid includes 180 households, with an installed photovoltaic capacity of 800kW. Through statistical analysis of the typical daily load of this microgrid in the summer of 2021, the maximum daily load was found to be 453.01kW, and the minimum daily load was 113.99kW. The typical daily load curve, corresponding time-of-use electricity price, and typical daily photovoltaic output of the microgrid are shown below. Figure 3 As shown, the electricity price is 0.68 yuan during normal hours, 1.06 yuan during peak hours, and 0.31 yuan during off-peak hours. The community has 100 electric vehicles, a user participation coefficient ξ = 0.8, and a percentage of idle vehicles p. B =0.2, the energy storage system capacity is 1000kW; the charging and discharging parameters of energy storage and electric vehicles are shown in Table 1.

[0146] Table 1 Simulation parameters of the community microgrid

[0147] Energy storage system charging efficiency 0.9 Electric vehicle charging mode slow charging Energy storage system discharge efficiency 0.9 Electric vehicle charging voltage 220V Minimum state of charge of energy storage system 0.1 Electric vehicle charging current 16A Maximum state of charge of energy storage system 0.9 Electric vehicle charging power 3.5kW Maximum charging power of energy storage system 200kW Electric vehicle battery capacity 24kWh Minimum charging power of energy storage system 200kW Electric vehicle charging and discharging efficiency 0.9

[0148] By analyzing the effect on load power regulation in the following four scenarios, the feasibility and effectiveness of the proposed electric vehicle-energy storage coordinated charging and discharging strategy are demonstrated.

[0149] Scenario 1: Energy storage is not considered and electric vehicles are charged and discharged in an unordered manner;

[0150] Scenario 2: Electric vehicles charge and discharge randomly, only considering the regulation of microgrid load power by energy storage charging and discharging;

[0151] Scenario 3: Without energy storage, only considering the regulation of microgrid load power by the orderly charging and discharging of electric vehicles;

[0152] Scenario 4: Adjusting microgrid load power by comprehensively considering energy storage and V2G (vehicle-to-grid) collaborative charging and discharging of electric vehicles.

[0153] According to the electric vehicle-energy storage coordinated control strategy proposed in this invention, after the energy storage and electric vehicle are connected to the grid in coordination, a three-objective optimization model is established to minimize the load peak-valley difference, minimize the mean square error of grid connection fluctuation, and minimize user expenditure costs. An improved non-dominated sorting genetic algorithm is used to solve the three-objective optimization problem and obtain the Pareto non-dominated solution set that satisfies the three objectives. Figure 4 This is the Pareto front of the three objective functions under the strategy proposed in this invention. Figure 5 The power allocation diagram under multi-objective optimization. Figure 6 To optimize the charging and discharging power and state of charge of an energy storage system under multiple objectives, Figure 7 This section describes the electricity purchase situation of community microgrids from the power grid in four scenarios.

[0154] The calculation results show that:

[0155] Depend on Figure 3 The peak-to-valley difference in community load is large. To ensure the good economic efficiency of the community microgrid system, a three-objective improved non-dominated sorting genetic algorithm is used for optimization calculations, with the preference coefficients for each sub-objective set as ρ1 = ρ2 = 0.4 and ρ3 = 0.2. Figure 4 The figure shows the Pareto front for multi-objective optimization in scenario 4. The solution results of the model converge significantly, and the overall point distribution is relatively uniform.

[0156] Figure 5 and Figure 6 The figures show the power allocation diagram and the charging / discharging power and state of charge of the energy storage system under multi-objective optimization in scenario 4, respectively. It can be seen that under the strategy formulated in this invention, and combined with... Figure 3 Analyzing the time-of-use pricing, the community significantly reduced its electricity purchases during peak hours, with electric vehicles primarily charging during off-peak hours. During peak hours, some electric vehicles could also discharge into the grid, thus reducing users' original expenses. Between 10:00 and 14:00, photovoltaic output was generally sufficient to meet user load demands. Therefore, despite this being peak electricity consumption, the energy storage system could still store some energy. This energy could be discharged during peak hours when photovoltaic output was low, and also provide energy for electric vehicle charging, thereby improving energy utilization within the community microgrid system, reducing the peak-to-valley difference in the load curve (purchased power), and lowering users' electricity costs. Table 2 shows the multi-objective optimization results for the four scenarios.

[0157] Table 2 Optimization results for multi-objective scenarios in four scenarios

[0158] Scene 1 698.33 671.69 2175.15 4678.03 Scene 2 577.26 572.75 1837.23 4091.80 Scene 3 500.65 500.65 1528.12 3303.41 Scene 4 440.47 368.47 1170.28 2271.74

[0159] Figure 7 This section describes the electricity purchase situation of community microgrids from the power grid in four scenarios.

[0160] A comparison of Scenario 1 and Scenario 2 shows that optimizing the charging and discharging power of energy storage smoothed out the load curve to some extent. The load peak-to-valley difference decreased from 671.69kW to 572.75kW, and the load mean square error, which mainly reflects the degree of load fluctuation, decreased from 2175.15kW. 2 It dropped to 1837.23kW 2 Because energy storage can be charged during off-peak hours and discharged during peak hours, it can effectively reduce the peak-valley load difference and reduce user expenses by 12.53%.

[0161] A comparison of scenarios one and three shows that after adopting time-of-use pricing to guide electric vehicles' charging and discharging, the load curve was smoothed out to some extent. The load peak-to-valley difference decreased from 671.69kW to 500.65kW, and the load mean square error, which mainly reflects the degree of load fluctuation, decreased from 2175.15kW. 2 It dropped to 1528.12kW 2 This is mainly because when electric vehicles without energy storage are charging and discharging in a disorderly manner, electric vehicle users return to the community one after another at night to engage in disorderly charging and discharging behavior. The load is concentrated in the evening when people return home, while the charging load is very low in the early morning and afternoon. This is mainly affected by the travel of users in the community and the electricity consumption behavior of residents. After the implementation of time-of-use pricing, the charging load of electric vehicles is delayed to the off-peak hours due to the incentive of low electricity prices. Compared with scenario 1, the user's expenditure is reduced by 29.38%.

[0162] Comparing Scenario 4 with the other three scenarios, when energy storage and electric vehicle (EV) V2G jointly participate in the optimization configuration, the ability to peak and valley load reduction is improved compared to when only one party participates in the optimization configuration. This is because their charging and discharging capabilities can be superimposed and complement each other when controlled collaboratively. For example, the charging and discharging behavior of EVs is largely constrained by user travel behavior. During this period, energy storage, with its more flexible scheduling, can be used for optimization. At the same time, the orderly charging and discharging behavior of EVs delays the overall charging load at night, thus achieving a peak and valley reduction capability that is difficult to achieve by energy storage systems alone. With the synergistic effect of energy storage and EVs, the peak-valley difference was reduced to 368.47kW, and the load mean square error reached 1528.12kW. 2 The load waveform is smoother, and the user's expenses are reduced by 51.44%, 44.48%, and 31.23% respectively compared to scenario one, scenario two, and scenario three.

[0163] The technical solution provided by this invention comprehensively considers the charging and discharging characteristics of electric vehicles and energy storage. In the photovoltaic-load-storage microgrid that takes into account the access of electric vehicles, it solves the multi-objective optimization problem of energy storage and electric vehicle V2G collaborative scheduling. This solution has certain advantages in effectively balancing the peak and valley loads of the system and reducing user expenses. It has important significance and reference value for solving the problem of optimal energy allocation.

Claims

1. A method for coordinated charging and discharging of electric vehicles and energy storage based on multi-objective optimization, characterized in that... Includes the following steps: Step 1: Construct a community microgrid system that includes photovoltaic power generation systems, energy storage systems, electric vehicles, and the electricity load of community users; Step 2: Use peak-valley electricity pricing strategy to control the charging and discharging behavior of electric vehicles, so that energy storage and electric vehicles can be connected to the grid in a coordinated manner, discharging during peak hours and charging during off-peak hours; Step 3: After energy storage and electric vehicles are connected to the grid in a coordinated manner, establish an objective optimization model that considers minimizing the load peak-valley difference, minimizing the mean square error of grid connection fluctuation, and minimizing user expenditure costs; Step 4: Solve the objective optimization model established in Step 3 using the improved non-dominated sorting genetic algorithm NSGA-II to obtain the set of Pareto non-dominated solutions that meet the requirements; Step 5: From the Pareto non-dominated solution set, select the optimal solution from the optimal solution set according to the fuzzy membership function, and use it as the planned output value for energy storage and electric vehicles; Step four includes the following steps: Step a: Based on the problem of coordinated charging and discharging control of community microgrid energy storage and electric vehicles, randomly generate... Given an individual, the algorithm's basic parameters are input, initial crossover and mutation probabilities are set, and each individual is encoded. A Cubic chaotic mapping strategy is employed, resulting in a more uniform initial population distribution and smaller differences between individuals. The Cubic chaotic mapping function is as follows: ； In the formula, For the first The location of each individual; , The chaos influence factor; the range of the Cubic mapping depends on , The value, At that time, its mapping state is a chaotic sequence; Indicates by the first The first individual chaotic mapping The location of each individual; Step b: Calculate the fitness value and constraint violation value of each individual in the population; Step c: Rank individuals based on their merits using an improved non-dominated ranking strategy and their crowding distance; Step d: Select dominant individuals as parent populations; Step e: Perform selection, crossover, and mutation operations on the parent generation to generate the offspring population; Step f: Merge the elite individuals selected by the elite strategy in the parent population with the offspring population generated in step e; Step g: Recalculate the individual fitness values ​​and constraint violation values ​​of the newly generated population; Step h: Determine if the maximum number of iterations has been reached. If yes, output the Pareto non-dominated solution set; otherwise, repeat step c. Step c includes an improved non-dominated sorting strategy: Introducing a new indicator for evaluating individual merits and demerits—Dominance Coupling Degree This indicates that the smaller the dominance coupling degree of an individual in a non-dominated set, the better its optimization performance in most sub-objectives of the multi-objective optimization problem. This individual has superior performance and is more likely to pass it on to its offspring. Its expression is: ； In the formula, individual It belongs to the non-dominant set; The number of sub-targets; and Representing the The maximum and minimum fitness values ​​of each sub-objective; Represents an individual fitness value; At this time, the individual There is a non-dominant hierarchy Dominant Coupling Degree Crowded distance Three attributes; For any two individuals ,like or or Then the individual Its performance is superior; Step e includes introducing an adaptive crossover mutation operator: ； In the formula, , These are the crossover probability and the mutation probability, respectively. , These represent the total number of evolutions and the current number of evolutions, respectively.

2. The electric vehicle-energy storage coordinated charging and discharging method based on multi-objective optimization according to claim 1, characterized in that: In step one: a: Photovoltaic power output is related to light intensity. We can use a beta distribution model to fit the distribution function of light intensity, and its probability density is: ； In the formula, Represents the gamma function; This represents the actual measured light intensity; Indicates the maximum light intensity; , These are the two shape parameters of the beta distribution; Its photovoltaic output power expression is: ； In the formula, This indicates the total output power of the photovoltaic cell array; Indicates the conversion efficiency of the photovoltaic cell module; This represents the total area of ​​the photovoltaic cell array; b: The relationship between the energy storage system's electrical capacity and its charging / discharging power is expressed as follows: ； In the formula, and They represent time periods respectively. and time period The amount of electricity stored in an energy storage system; Indicates time period The power that the energy storage system can output; For time period and time period The time interval between; The charging and discharging process of an energy storage system is as follows: (1) Discharge process: Energy storage system in the The sky The discharge power per hour is: ； In the formula, This refers to the rated discharge power of the energy storage system. This represents the maximum capacity of the energy storage system. For energy storage systems in the first Heavenly Energy storage capacity for a given time period; Due to its own energy storage capacity, its expression is: ； In the formula, Indicates the energy storage system in the first Heavenly Energy storage capacity for a given time period; This represents the minimum capacity of the energy storage system. The ideal capacity of an energy storage system is expressed as: ； In the formula, The discharge efficiency of the energy storage system; (2) Charging process: Energy storage system in the sky The charging power per hour is: ； In the formula, Rated charging power for the energy storage system; ； In the formula, The charging efficiency of the energy storage system; Indicates the energy storage system in the first Heavenly Charging power during a given period.

3. The electric vehicle-energy storage coordinated charging and discharging method based on multi-objective optimization according to claim 1, characterized in that: In step two: due to the behavioral characteristics of electric vehicle owners, the discharge process is divided into daytime peak hours and evening peak hours, and two discharge strategies are formulated: (1).No. Daytime peak hours The actual discharge power of all electric vehicles participating in the time-of-use pricing response within the hour is: ； In the formula, This indicates the number of electric vehicles participating in discharge during peak daytime hours; Represents the response to time-of-use pricing. Electric vehicles; For the first The expression for the daytime peak discharge power of a Taiwanese electric vehicle is as follows: ； In the formula, For discharge efficiency; The electric vehicle is in a fully charged state; This represents the minimum state of charge during discharge. For electric vehicle capacity; ； In the formula, The proportion of electric vehicles that are idle; The user participation coefficient represents the proportion of electric vehicle owners who respond to the time-of-use pricing strategy out of the total number of electric vehicle owners; The total number of electric vehicles; This refers to the number of peak daytime periods; (2).No. evening rush hour The actual discharge power of all electric vehicles participating in the time-of-use pricing response within the hour is: ； in: ； ； ； In the formula, This refers to the number of evening peak hours; Electric vehicles that are in use and responding to time-of-use pricing; Let be the probability function of the electric vehicle's return time; This refers to the number of electric vehicles participating in discharge during the evening peak period; Electric vehicle charging strategies: No. Tianping Valley period The amount of charge a car receives in one hour is: ； In the formula, Pinggu period Hourly charging load.

4. The electric vehicle-energy storage coordinated charging and discharging method based on multi-objective optimization according to claim 1, characterized in that: In step two: by setting parameters By adjusting the proportions of energy storage systems and electric vehicles connected to the grid during peak and off-peak hours respectively, the model for their coordinated grid-connected power output is as follows: ； In the formula, For the first Heavenly Time-of-use energy storage and the discharge power of electric vehicles; For the first Heavenly Time-of-use energy storage and charging power for electric vehicles; Indicates the first Heavenly Discharge power of the time-limited energy storage system; Indicates the first Heavenly The discharge power of electric vehicles participating in time-of-use pricing during specific time periods; Indicates the first Heavenly Charging power of time-limited energy storage systems; Indicates the first Heavenly The charging power of electric vehicles during a given time period; Here, it is assumed that the total power provided by all electric vehicles participating in the time-of-use pricing response during peak hours is... The total power that all energy storage devices can provide is Then we can obtain: ； Similarly, it can be concluded that during the Pinggu period... The possible values ​​are as follows: Assuming that the total power required for charging of all electric vehicles participating in time-of-use pricing during off-peak hours is The total power required by all energy storage devices is Then we can obtain: ； After energy storage and electric vehicles are connected to the grid in a coordinated manner, the first Heavenly Time-series load during the period for: ； In the formula, , The first Heavenly The original load power and photovoltaic output power during the time period.

5. The electric vehicle-energy storage coordinated charging and discharging method based on multi-objective optimization according to claim 1, characterized in that: In step three: with the goal of achieving the best peak-shaving and valley-filling effect on the load curve after the coordinated grid connection of energy storage and electric vehicles, an optimization objective function considering the minimum load peak-valley difference is established. And the optimization objective function that considers minimizing the mean square error of grid connection fluctuations as follows: ； ； In the formula, , The first The maximum and minimum time-series load power of the day; The number of time periods in a day; Establish an optimization objective function that minimizes user spending. : ； In the formula, Electricity prices for charging at different times; These represent the non-charging and charging states, respectively.

6. The electric vehicle-energy storage coordinated charging and discharging method based on multi-objective optimization according to claim 5, characterized in that: The constraints on the above objective function are: ① Photovoltaic output constraints: ； , These represent the upper and lower limits of photovoltaic power output, respectively. Indicates the first Heavenly The output power of photovoltaics during a given period; ② Capacity constraints of energy storage: ； ③ Safety constraints for electric vehicle batteries: ； Indicates the lower limit of the battery's state of charge; Indicates the first Heavenly Time period The state of charge of an electric vehicle; Indicates the upper limit of the battery's state of charge; ④ Power balance constraints: ； In the formula, For the first Heavenly Power supplied by the power grid during a given time period; Indicates the first Heavenly Time-series load during a given period.

7. The electric vehicle-energy storage coordinated charging and discharging method based on multi-objective optimization according to claim 1, characterized in that: Step five is specifically as follows: First, iterate through the Pareto solution set and calculate the solution of the first Pareto solution. The solution of the first Membership degree of each objective function: ； , They represent the first The minimum and maximum fitness values ​​of each objective function; Decision-makers adjust preference coefficients, i.e., the weights of each objective, based on load conditions. A weighted method is used to obtain the membership weights for the optimal solution. The Pareto solution corresponding to the maximum value is the optimal solution to the problem. The calculation formula is as follows: ； In the formula, For the first The preference coefficients of each sub-objective function; Population size; The membership function value; This represents the total number of objective functions; Indicates the first One objective function.

Images