Ordered charging and discharging control method for electric vehicle based on improved differential evolution algorithm

Abstract

The invention discloses an electric vehicle ordered charging and discharging control method based on an improved differential evolution algorithm. The method comprises the following steps: constructing a probability model of an electric vehicle charging behavior through a Monte Carlo method; demand analysis and electric vehicle dispatching feasibility judgment are carried out by acquiring power grid information and charging electric vehicle information; considering the power grid operation safety risk and the user credit rating, and formulating a dynamic electricity price; establishing a charging and discharging control model by taking minimization of the comprehensive cost of the user and the construction cost of the power grid as a target; and an improved differential evolution algorithm is adopted to solve the model to obtain the optimal electricity price and the optimal charging and discharging strategy, so that the operation efficiency and the economical efficiency of the power system are improved. The method effectively overcomes the defects of flexibility and cost of an existing electricity price strategy and the defects of dynamic electricity price making in the aspect of user credit, can effectively cope with power grid load fluctuation, and promotes coordinated development of the electric vehicle and the power grid.

Description

Technical Field

[0001] The present invention belongs to the field of electric vehicle optimization scheduling, and specifically is an electric vehicle orderly charging and discharging control method based on an improved differential evolution algorithm. Background Art

[0002] In recent years, the rapid development of the electric vehicle (EV) industry is profoundly affecting the global energy structure and environmental protection. However, with the surge in the number of electric vehicles, the load pressure on the power grid has also increased, posing new challenges to the operational safety and economy of the power system. The current electricity price strategy and charging and discharging control methods are unable to cope with these challenges. Traditional electricity price strategies often lack sufficient flexibility and adaptability, while existing charging and discharging control schemes fail to effectively cope with the drastic fluctuations in grid load and the economic benefits of users, which makes the power system face greater management challenges in the process of popularizing electric vehicles. Therefore, it is urgent to develop more intelligent and dynamic electricity prices and control strategies to achieve efficient coordination and optimized operation of electric vehicles and power grids.

[0003] Electric vehicle loads are highly flexible and dispatchable, but existing time-of-use electricity pricing strategies have obvious shortcomings in dealing with grid load fluctuations. Due to the lack of a dynamic adjustment mechanism, these strategies cannot reflect the actual operating status of the grid in real time, resulting in waste of power resources and grid safety hazards. In addition, the existing dynamic electricity price formulation methods fail to effectively combine user credit ratings and grid safety requirements, resulting in deficiencies in the reliability and fairness of the electricity price mechanism. At the same time, traditional charging and discharging control methods often ignore the comprehensive consideration of grid construction and operating costs, and fail to fully integrate the charging and discharging needs of electric vehicles with the overall operation of the grid, thereby limiting the optimization of charging and discharging strategies and failing to maximize the economic benefits and operating efficiency of the system. Summary of the invention

[0004] The present invention provides an orderly charging and discharging control method for electric vehicles based on an improved differential evolution algorithm, which is applied to dynamic electricity price formulation and optimization of charging and discharging control strategies. It effectively overcomes the deficiencies of existing electricity price strategies in terms of flexibility and cost, as well as the defects of dynamic electricity price formulation in terms of user credit, and can effectively cope with grid load fluctuations and promote the coordinated development of electric vehicles and grids.

[0005] A method for orderly charging and discharging control of electric vehicles based on an improved differential evolution algorithm comprises the following steps:

[0006] Step 1. Using the Monte Carlo method, a probability model of electric vehicle charging behavior including charging start time, charging end time, daily mileage and charging time is established;

[0007] Step 2. By acquiring information of the power grid and the charging electric vehicles, the smart charging pile performs demand analysis and determines the feasibility of electric vehicle scheduling;

[0008] Step 3: By calculating the current grid security risk value and dividing users into different credit levels, we can construct credit assessment coefficients for different users. On this basis, we can formulate a flexible dynamic electricity price strategy.

[0009] Step 4: Considering the comprehensive charging cost of electric vehicle users and the grid construction cost, and based on the charging and discharging status, state of charge and power limit of electric vehicles, an electric vehicle charging and discharging strategy optimization scheduling model is constructed;

[0010] Step 5: Use the improved differential evolution algorithm to solve the optimization scheduling model, obtain the optimal dynamic electricity price and the corresponding orderly charging and discharging control strategy, and execute it by the smart charging pile.

[0011] The present invention adopts a dynamic electricity price formulation strategy, combines the grid load fluctuation risk coefficient and the user credit grading system, and formulates a flexible and economical electricity price strategy. This strategy not only improves the grid's response speed to load changes, but also achieves differentiation of electricity prices through the credit grading mechanism, so that users with different credit levels receive corresponding treatment during the electricity price adjustment process. This not only effectively reduces the user's charging cost, but also reduces the peak-to-valley difference and load fluctuation, thereby improving the safety and economy of grid operation.

[0012] The present invention optimizes the charging and discharging strategy by comprehensively considering the grid construction cost and the comprehensive cost of users. The present invention reduces the grid construction and operation expenses while maximizing the economic benefits of users. Comprehensive analysis of the grid construction cost and the user charging cost effectively reduces the demand for grid expansion, improves the economic benefits and operating efficiency of the system, and at the same time enhances the safety and stability of the grid.

[0013] The improved differential evolution algorithm used in the present invention not only improves the solution efficiency, but also enhances the global search capability and the adaptability to the dynamic environment, so as to more accurately solve the optimal dynamic electricity price and orderly charging and discharging control strategy, further optimize the electric vehicle scheduling effect, and ensure the stability and safety of the power grid operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 is a flow chart of the electric vehicle dispatching method according to the present invention;

[0015] Figure 2 Develop comparison charts for dynamic electricity prices;

[0016] Figure 3 It is a comparison chart of total load and foundation load in original state;

[0017] Figure 4 The load curve diagram after the strategy optimization of the present invention when charging an EV. DETAILED DESCRIPTION

[0018] like Figure 1 As shown, a method for orderly charging and discharging control of electric vehicles based on an improved differential evolution algorithm comprises the following steps:

[0019] Step 1. Using the Monte Carlo method, a probability model of electric vehicle charging behavior including charging start time, charging end time, daily mileage and charging time is established;

[0020] The probability distribution model of electric vehicle charging behavior consists of a charging start time model, a charging end time model, a daily mileage model, and a charging duration model;

[0021] The charging start time model is:

[0022]

[0023] Where: t a Indicates the charging start time; μ a Indicates the charging start time t a The mean of a represents the standard deviation of the charging start time ta; f(ta) represents the probability function of the charging start time ta.

[0024] The charging end time model is:

[0025]

[0026] Where: t i Indicates the end time of charging; μ1 indicates the end time of charging t i The mean of a Indicates the charging end time t i The standard deviation of f(t i ) represents the charging end time t i The probability function of .

[0027] The daily mileage conforms to the log-normal distribution f d (x) is:

[0028]

[0029] Where: x represents daily mileage; μ l represents the mean of mileage; σ l represents the standard deviation of mileage; lnx represents the natural logarithm of mileage.

[0030] Combined with the travel characteristics of EV users, the time required to schedule EVi charging can be calculated. i :

[0031]

[0032] Where: Ti is the charging time of the i-th electric vehicle; S e is the battery capacity of electric vehicles; SOC d The state of charge of the battery at the end of charging an electric vehicle; SOC i The state of charge of the battery at the beginning of charging the electric vehicle; P i is the charging power of the electric vehicle; η is the charging efficiency of the electric vehicle.

[0033] Step 2. By acquiring information of the power grid and the charging electric vehicles, the smart charging pile performs demand analysis and determines the feasibility of electric vehicle scheduling;

[0034] By obtaining information from the power grid and the charging electric vehicle, the information includes: expected departure time, expected power and dispatch willingness, remaining state of charge SOC (State Of Charge), EV battery capacity, and basic load information of the vehicle. The demand analysis described in it is:

[0035]

[0036] Where: I = 1 indicates that the electric vehicle has the intention to dispatch, and I = 0 indicates that the electric vehicle has no intention to dispatch; Feasibility judgment of electric vehicle dispatch:

[0037]

[0038] Where: T i Indicates the time required to charge an electric vehicle, T L represents the time when the electric vehicle is off the grid, T a Indicates the time when the electric vehicle is connected to the grid.

[0039] Step 3: By calculating the current grid security risk value and dividing users into different credit levels, we can construct credit assessment coefficients for different users. On this basis, we can formulate a flexible dynamic electricity price strategy.

[0040] The grid security risk value and user credit rating are taken into consideration to formulate a charging and discharging price guidance curve for electric vehicles that meets the grid's own interests and send it to lower-level electric vehicle users. Dynamic electricity prices are formulated on this basis.

[0041] Consider the power grid security risk value as follows:

[0042]

[0043] Where P(t) is the total power demand in the region at time t; dt is a small time increment used to accumulate the total power demand of P(t) over the entire time interval [0,T]; T is the total time of a day;

[0044] Considering the user credit classification: first level credit, S = 1; medium credit is second level, S = 2; low credit is third level, S = 3; the user credit evaluation coefficient is:

[0045] Formulate dynamic electricity prices:

[0046] When P(t)≥P thresh Time: Electricity price formula:

[0047]

[0048] When P(t) <P thresh Time: Electricity price formula:

[0049]

[0050] Where: P thresh is the grid security risk value; C0 is the basic electricity price; β is the price growth coefficient; n is the value of the sensitivity of electricity price to load; α is the electricity price reduction coefficient, which controls the decline of electricity price when the load is low; γ is the attenuation rate of low load. S is the user's current credit score, Smax is the maximum credit score, and k2 is the credit discount coefficient.

[0051] The dynamic electricity price strategy of the present invention, such as Figure 2 The strategy shown in the figure adjusts the electricity price according to the real-time load of the power grid and the user's credit rating. By encouraging users to charge during periods with lower electricity prices, the peak load of the power grid can be effectively reduced. Users with different credit ratings also enjoy different electricity prices: users with high credit ratings have relatively low charging prices, while users with low credit ratings have relatively high charging prices. The electricity price curve can be flexibly tracked according to the fluctuations of the power grid, so that electric vehicle users can adjust their charging and discharging behaviors according to the electricity price, thereby improving the safety of the power grid.

[0052] The charging and discharging strategy optimization model of the present invention takes into account the comprehensive charging cost of electric vehicle users and the construction cost of the power grid, and its objective function is:

[0053] min{C chg +C Dis +C Bloss +C grid}

[0054] Where: C chg is the charging cost of electric vehicles, C Dis The discharge reward for electric vehicles, CBloss is the battery aging cost of electric vehicles during charging, C grid The cost of power grid construction.

[0055]

[0056] Where: is the unit electricity price for charging an electric vehicle at time t, The real-time power of charging the electric vehicle at time t; is the charging status of the electric vehicle at time t; T c is the discharge time of the electric vehicle; is the unit electricity price of electric vehicle discharged at time t, is the real-time power discharged by the electric vehicle at time t, is the discharge state of the electric vehicle at time t, T d is the discharge time of the electric vehicle, C B is the investment cost of electric vehicle batteries, C ev is the battery capacity of electric vehicles, D d is the discharge depth of the electric vehicle battery, L c The cycle life of electric vehicle batteries; is the charging power of the electric vehicle at time t, m is the investment cost of the unit generator set, Pw is the power of the power grid through the generator standby unit for auxiliary peak shaving; n is the number of years used; i is the annual interest rate; is the coefficient for converting the construction investment cost of the generating set capacity into the equivalent annual value; ε is the spinning reserve rate adopted when constructing the generating set; δ is the labor maintenance cost and construction investment cost required after the construction of the unit load unit.

[0057] Construct a constrained charging and discharging strategy optimization scheduling model;

[0058] The objective function of the charging and discharging strategy formulation stage is subject to the constraints of the electric vehicle charging and discharging state, the electric vehicle state of charge, and the electric vehicle charging and discharging power. The constraints are established:

[0059] Electric vehicle charging and discharging state constraints:

[0060]

[0061] Electric vehicle state of charge constraints:

[0062] SOC min ≤SOC d ≤SOC max

[0063] SOC min ≤SOC i ≤SOCmax

[0064]

[0065] Where: ΔT represents the time interval between the end of charging and the start of charging, T β T is the end time of charge and discharge; c is the start time of charging and discharging; SOC i SOC is the charge capacity of the electric vehicle when it is connected to the grid. d SOC is the charge capacity of an electric vehicle when it leaves the grid; min With SOC max They represent the minimum and maximum states of charge of electric vehicle batteries, respectively. and The efficiency of charging and discharging electric vehicles, The rated capacity of the electric vehicle battery;

[0066] Electric vehicle charging and discharging power constraints:

[0067]

[0068] Where: is the maximum charging power of electric vehicles, is the maximum discharge power of the electric vehicle.

[0069] Step 4: Considering the comprehensive charging cost of electric vehicle users and the grid construction cost, and based on the charging and discharging status, state of charge and power limit of electric vehicles, an electric vehicle charging and discharging strategy optimization scheduling model is constructed;

[0070] Step 5: Use the improved differential evolution algorithm to solve the optimization scheduling model, obtain the optimal dynamic electricity price and the corresponding orderly charging and discharging control strategy, and execute it by the smart charging pile.

[0071] The steps of the improved differential evolution algorithm include population initialization, cluster analysis, mutation operation, crossover operation, selection operation, and information exchange between subgroups.

[0072] Step 6.1: Population initialization: Randomly generate M individuals that meet the constraints in n-dimensional space:

[0073]

[0074] Where: x ij (0) is the initial value of the jth parameter of the ith individual at time t = 0; and are the upper and lower bounds of the jth parameter, respectively, rand ij (0,1) is a random number between [0,1].

[0075] Step 6.2: Cluster analysis (using density peak clustering algorithm): Perform cluster analysis on the population and divide the population into K subgroups. Specific steps:

[0076] Step 6.2.1. Calculate the distance matrix: Calculate the Euclidean distance between individuals in the population and construct the distance matrix D;

[0077] D ij =||x i -x j ||

[0078] Step 6.2.2. Calculate the local density: For each individual x i , calculate its local density ρ i , k-nearest neighbors are usually used to define local density:

[0079]

[0080] Where: N i is an individual x i is the set of k-nearest neighbors of , and σ is the scale parameter of the Gaussian kernel.

[0081] Step 6.2.3. Calculate the distance peak: For each individual x i , calculate its distance peak δ i :

[0082] δ i =min{dist ij |ρ j >ρ i}

[0083] Where dist ij Represents individual x i to x j The distance; ρ i For individual x i The local density of j For individual x j The local density of .

[0084] Step 6.2.4. Select the peak point: According to the local density ρ i and the distance peak δ i , select the density peak point and use it as the cluster center: ρ i and δ i Maximum, select x i As the peak point, the individual is also selected as the cluster center.

[0085] Step 6.2.5. Clustering: Assign each individual to the cluster with the density peak point closest to it.

[0086] Step 6.3: Mutation operation: Randomly select 3 individuals x from the subpopulation p1 、x p2 and x p3 And p1≠p2≠p3.

[0087] h ij (t+1)=x p1j (t)+F(x p2j (t)-x p3j (t))

[0088] Where: h ij (t+1) is the value of the jth parameter of the ith individual after the mutation operation; x p1j (t) is the value of the jth parameter of the first randomly selected individual in the current generation; x p2j (t)-x p3j (t) is the variation term. F is the variation factor, p1, p2, and p3 are random integers, which are the serial numbers of individuals in the population.

[0089] Step 6.4: Crossover operation: In order to increase the diversity of the population, a crossover operation is required to generate new individuals.

[0090]

[0091] Where: v ij (t+1) is the candidate value of the jth parameter of the i-th individual after the crossover operation. ; x ij (t) is the value of the jth parameter of the ith individual in the current generation; randl ij is a random number between [0,1], CR is the crossover probability, CR∈[0,1].

[0092] Step 6.5: Select an operation: To determine x ij (t) Whether they can become members of the next generation, v ij (t+1) and x ij (t) Compare the evaluation functions

[0093]

[0094] Among them, x ij (t+1) is the value of the jth parameter of the ith individual in the next generation; f(v i1 (t+1),...,v in (t+1)) is the fitness function value of the i-th individual in the new generation, where v i1 (t+1),...,v in (t+1) is the candidate value of all parameters of the individual;

[0095] Step 6.6: Information exchange between subgroups: Every several generations, allow subgroups to exchange information to facilitate global search. The specific steps are as follows:

[0096] Step 6.6.1 Select the best individual: Select each subgroup C k The optimal individual x best k(t), and the global optimal individual x global best(t).

[0097] Step 6.6.2 Generate new individuals: Generate new individuals using the global optimal individual and the optimal individuals of other subgroups:

[0098]

[0099] in, is the new value of the jth parameter of the ith individual after information exchange. ; x best k(t) is the best individual of the Kth subgroup in the current generation. ; is the optimal individual of other subgroups; F is the variation factor, which controls the influence of the global optimal individual on the generation of new individuals;

[0100] Step 6.6.3 Replacement operation: Compare the newly generated individuals with the current individuals and select the ones with better fitness as the next generation members:

[0101]

[0102] Step 6.7: Termination condition: Repeat the above steps until the preset termination condition is met (the global optimal value no longer changes).

[0103] Step 6.8: Output results: After the algorithm terminates, output the global optimal solution

[0104] The optimal charging and discharging strategy of electric vehicles is solved by using an improved differential evolution algorithm. The strategy of the present invention takes into account the load fluctuation of the power grid, the credit rating of users, and the security risk value of the power grid to achieve the coordinated development of electric vehicles and the power grid. Figure 3, is the load condition of the power grid when no dispatching strategy is implemented. The basic load curve is obtained by collecting historical data of the power grid when there is no impact of electric vehicle charging load. This represents the traditional load mode of the power grid, which is usually relatively stable and does not take into account the additional load brought by electric vehicle charging. On the basis of the basic load curve, the data of electric vehicle charging load is added to obtain the total load curve. This curve reflects the actual impact of electric vehicle charging activities on the power grid load, especially during peak hours, when the charging behavior of electric vehicles will significantly increase the load pressure of the power grid, thus posing a challenge to the safety and stability of the power grid. This concentrated charging demand may cause local power grid overload, increase the risk of power outages, and adversely affect the dynamic balance of the power grid. Therefore, optimizing the charging strategy of electric vehicles to match it with the load characteristics of the power grid is crucial to maintaining the safe operation of the power grid.

[0105] like Figure 4 As shown, the electric vehicle scheduling method of the present invention is applied to execute the optimal charging and discharging strategy through the intelligent charging pile. The load curve in the disordered charging state in the figure is obtained by collecting the impact data of the random charging behavior of electric vehicles on the grid load. This curve reveals the grid load caused by the natural distribution of electric vehicle charging demand when there is no scheduling intervention, especially during peak hours, the load peak is particularly obvious. By implementing the scheduling strategy of the present invention, the intelligent charging pile optimizes the charging behavior of the electric vehicle to form a load curve in an orderly charging state. This curve shows that under the guidance of the optimal charging and discharging strategy, the charging demand of the electric vehicle is effectively redistributed to the period when the grid load is lower, thereby reducing the impact on the grid, showing the balanced distribution of the grid load after the implementation of the electric vehicle scheduling strategy, significantly reducing the peak-to-valley difference, and making the grid operation more stable and efficient.

Claims

1. An orderly charging and discharging control method for electric vehicles based on an improved differential evolution algorithm, characterized in that: The following steps are involved: Step 1. Using the Monte Carlo method, a probability model of electric vehicle charging behavior including charging start time, charging end time, daily mileage and charging time is established; Step 2. By acquiring information of the power grid and the charging electric vehicles, the smart charging pile performs demand analysis and determines the feasibility of electric vehicle scheduling; Step 3: By calculating the current grid security risk value and dividing users into different credit levels, we can construct credit assessment coefficients for different users. On this basis, we can formulate a flexible dynamic electricity price strategy. Step 4: Considering the comprehensive charging cost of electric vehicle users and the grid construction cost, and based on the charging and discharging status, state of charge and power limit of electric vehicles, an electric vehicle charging and discharging strategy optimization scheduling model is constructed; Step 5: Use the improved differential evolution algorithm to solve the optimization scheduling model, obtain the optimal dynamic electricity price and the corresponding orderly charging and discharging control strategy, and execute it by the smart charging pile.

2. The method for orderly charging and discharging control of electric vehicles based on an improved differential evolution algorithm according to claim 1, characterized in that: In step 1, the probability distribution model of the electric vehicle charging behavior is composed of a charging start time model, a charging end time model, a daily mileage model and a charging duration model; The charging start time model is: Where: t a Indicates the charging start time; μ a Indicates the charging start time t a The mean of a Indicates the charging start time t a The standard deviation of f(t a ) represents the charging start time t a The probability function of The charging end time model is: Where: t i Indicates the charging end time; u i Indicates the charging end time t i The mean of a Indicates the charging end time t i The standard deviation of f(t i ) represents the charging end time t i The probability function of The daily mileage conforms to the log-normal distribution f d (x) is: Where: x represents daily mileage; μ l represents the mean of mileage; σ l represents the standard deviation of mileage; lnx represents the natural logarithm of mileage; Combined with the travel characteristics of EV users, the time required to schedule EVi charging is calculated. i : Where: T i is the charging time of the i-th electric car; S e is the battery capacity of electric vehicles; SOC d The state of charge of the battery at the end of charging an electric vehicle; SOC i The state of charge of the battery at the beginning of charging the electric vehicle; P i is the charging power of the electric vehicle; η is the charging efficiency of the electric vehicle.

3. According to claim 1, an electric vehicle orderly charging and discharging control method based on an improved differential evolution algorithm is characterized in that: In the step 2, the grid information and the information of the charging electric vehicle are obtained, including: the expected departure time, the expected power and the dispatchable willingness, the remaining state of charge SOC, the EV battery capacity, and the basic load information of the vehicle; the demand analysis is as follows: Where: I = 1 indicates that the electric vehicle has the intention to dispatch, and I = 0 indicates that the electric vehicle has no intention to dispatch; Feasibility determination of electric vehicle dispatching: Where: T i Indicates the time required to charge an electric vehicle, T L represents the time when the electric vehicle is off the grid, T a Indicates the time when the electric vehicle is connected to the grid.

4. According to claim 1, an electric vehicle orderly charging and discharging control method based on an improved differential evolution algorithm is characterized in that: In step 3, the grid security risk value and user credit rating are considered to formulate a charging and discharging price guidance curve for electric vehicles that meets the interests of the grid itself and send it to lower-level electric vehicle users, and a dynamic electricity price is formulated on this basis; Consider the power grid security risk value as follows: Where P(t) is the total power demand in the region at time t; dt is a small time increment used to accumulate the total power demand of P(t) over the entire time interval [0,T]; T is the total time of a day; Consider the user credit classification: first level credit, S = 1; medium credit is second level, S = 2; low credit is third level, S = 3; Its user credit evaluation coefficient is: Formulate dynamic electricity prices: When P(t)≥P thresh Time: Electricity price formula: When P(t) <P thresh Time: Electricity price formula: Where: P thresh is the grid security risk value; C0 is the basic electricity price; β is the price growth coefficient; n is the value of the sensitivity of electricity price to load; α is the electricity price reduction coefficient, which controls the decline of electricity price under low load; γ is the attenuation rate of low load; S is the user's current credit score, S max is the maximum value of the credit score, and k2 is the credit discount factor.

5. According to claim 1, an electric vehicle orderly charging and discharging control method based on an improved differential evolution algorithm is characterized in that: In step 3, the charging and discharging strategy optimization model is: Considering the comprehensive charging cost of electric vehicle users and the construction cost of the power grid, the objective function is: my{C chg +C Dis +C Bloss +C grid } Where: C chg is the charging cost of electric vehicles, C Dis The discharge reward for electric vehicles, C Bloss is the battery aging cost of electric vehicles during charging, C grid The cost of building the power grid; Where: is the unit electricity price for charging an electric vehicle at time t, The real-time power of charging the electric vehicle at time t; is the charging status of the electric vehicle at time t; T c The discharge time of the electric vehicle; is the unit electricity price of electric vehicle discharged at time t, is the real-time power discharged by the electric vehicle at time t, is the discharge state of the electric vehicle at time t, T d is the discharge time of the electric vehicle, C B is the investment cost of electric vehicle batteries, C ev is the battery capacity of electric vehicles, D d is the discharge depth of the electric vehicle battery, L c is the cycle life of the electric vehicle battery; m is the investment cost of the unit generator set; Pw is the amount of electricity used by the power grid to assist in peak load reduction through the generator standby set; n is the number of years in use; j is the annual interest rate; is the coefficient for converting the construction investment cost of the generating set capacity into the equivalent annual value; ε is the spinning reserve rate adopted when constructing the generating set; δ is the labor maintenance cost and construction investment cost required after the construction of the unit load unit.

6. The method for orderly charging and discharging control of electric vehicles based on an improved differential evolution algorithm according to claim 5 is characterized in that: Construct a constrained charging and discharging strategy optimization scheduling model; the objective function of the charging and discharging strategy formulation stage is subject to the constraints of the electric vehicle charging and discharging state, the electric vehicle charge state and the electric vehicle charging and discharging power, and establish the constraints: Electric vehicle charging and discharging state constraints: Electric vehicle state of charge constraints: SOC min ≤SOC d ≤SOC max SOC min ≤SOC i ≤SOC max Where: ΔT represents the time interval between the end of charge and discharge and the start of charge and discharge, T β T is the end time of charge and discharge; c is the start time of charging and discharging; SOC i SOC is the charge capacity of the electric vehicle when it is connected to the grid. d SOC is the charge capacity of an electric vehicle when it leaves the grid; min With SOC max They represent the minimum and maximum states of charge of electric vehicle batteries, respectively. and The efficiency of charging and discharging electric vehicles, is the rated capacity of the electric vehicle battery, SOC l Indicates the battery's state of charge after a period of charging and discharging; Electric vehicle charging and discharging power constraints: Where: is the maximum charging power of electric vehicles, is the maximum discharge power of the electric vehicle.

7. The method for orderly charging and discharging control of electric vehicles based on an improved differential evolution algorithm according to claim 1 is characterized in that: The improved differential evolution algorithm is: The steps of the improved differential evolution algorithm include population initialization, cluster analysis, mutation operation, crossover operation, selection operation, and information exchange between subgroups; Step 6.1: Population initialization: Randomly generate M individuals that meet the constraints in the multidimensional space: Where: x ij (0) is the initial value of the jth parameter of the ith individual at time t = 0; and are the upper and lower bounds of the jth parameter, respectively, rand ij (0,1) is a random number between [0,1]; Step 6.2: Cluster analysis: Perform cluster analysis on the population and divide the population into K subgroups; specific steps: Step 6.2.1: Calculate the distance matrix: Calculate the Euclidean distance between individuals in the population and construct the distance matrix D ij ; D ij =||x i -x j || Step 6.2.

2. Calculate the local density: For each individual x i , calculate its local density ρ i , k-nearest neighbors are usually used to define local density: Where: N i is an individual x i The k-nearest neighbor set of , ο is the scale parameter of the Gaussian kernel; Step 6.2.

3. Calculate the distance peak: For each individual x i , calculate its distance peak δ i : d i =min{dist ij ∣r j >r j } Where dist ij Represents individual x i to x j The distance; ρ i For individual x i The local density of j For individual x j The local density of Step 6.2.

4. Select the peak point: According to the local density ρ i and the distance peak δ i , select the density peak point and use it as the cluster center: ρ i and δ i Maximum, select x i As the peak point, the individual is also selected as the cluster center; Step 6.2.

5. Clustering: Assign each individual to the cluster with the nearest density peak point; Step 3: Mutation operation: Randomly select 3 individuals x from the subpopulation p1 、x p2 and x p3 And p1≠p2≠p3; h ij (t+1)=x p1j (t)+F(x p2j (t)-x p3j (t)) Where: h ij (t+1) is the value of the jth parameter of the ith individual after the mutation operation; x p1j (t) is the value of the jth parameter of the first randomly selected individual in the current generation; x p2j (t)-x p3j (t) is the variation term; F is the variation factor, p1, p2, p3 represent random integers, and are the serial numbers of individuals in the population; Step 6.4: Crossover operation: In order to increase the diversity of the population, a crossover operation is required to generate new individuals; Where: v ij (t+1) is the candidate value of the jth parameter of the i-th individual after the crossover operation. ; x ij (t) is the value of the jth parameter of the ith individual in the current generation; randl ij is a random number between [0,1], CR is the crossover probability, CR∈0,1]; Step 6.5: Select an operation: To determine x ij (t) Whether they can become members of the next generation, v ij (t+1) and x ij (t) comparing the evaluation functions; Among them, x ij (t+1) is the value of the jth parameter of the ith individual in the next generation; f(v i1 (t+1),...,v in (t+1)) is the fitness function value of the i-th individual in the new generation, where v i1 (t+1),...,v in (t+1) is the candidate value of all parameters of the individual; Step 6.6: Information exchange between subgroups: Every several generations, allow information exchange between subgroups to promote global search; the specific steps are as follows: Step 6.6.1: Select the best individual: Select each subgroup C k The optimal individual x best k(t), and the global optimal individual x global best(t); Step 6.6.2 Generate new individuals: Generate new individuals using the global optimal individual and the optimal individuals of other subgroups: in, is the new value of the jth parameter of the ith individual after information exchange. ; x best k(t) is the best individual of the Kth subgroup in the current generation. ; is the optimal individual of other subgroups; F is the variation factor, which controls the influence of the global optimal individual on the generation of new individuals; Step 6.6.3 Replacement operation: Compare the newly generated individuals with the current individuals and select the ones with better fitness as the next generation members: Step 6.7: Termination condition: Repeat the above steps until the maximum number of iterations is met; Step 6.8: Output results: After the algorithm terminates, output the global optimal solution

Images