A county photovoltaic district energy storage capacity optimization configuration method containing an electric vehicle

Abstract

The application discloses a kind of county photovoltaic substation energy storage capacity optimization configuration methods containing electric vehicle, it includes the following steps: S1: establishing the mathematical model of each elastic resource in substation and schedulable potential model, including: distributed photovoltaic power generation model, substation load model, energy storage adjustable potential model, electric vehicle charging and discharging model and electric vehicle schedulable potential model;S2: establishing energy storage capacity double-layer optimization configuration model, wherein, upper layer is capacity optimization configuration model, lower layer is optimization scheduling model;S3: upper layer model with energy storage installed capacity as decision variable, considering energy storage installed constraint and so on constraint condition, establish peak clipping and valley filling objective function and new energy consumption objective function;S4: lower layer model with electric vehicle charging and discharging power and energy storage charging and discharging power as decision variable, establish the objective function of minimum substation operating cost.The application has the characteristics of good convergence, high concentration of solution set, and good feasibility in engineering practice.

Description

TECHNICAL FIELD

[0001] The application relates to a photovoltaic substation energy storage configuration method, in particular to a county photovoltaic substation energy storage capacity optimization configuration method containing electric vehicles. BACKGROUND

[0002] Under the background of a new power system, distributed photovoltaic has the advantages of low construction cost, high income, and no need to build a distribution station, and is the main direction of photovoltaic market development. A substation is a power supply area managed by a transformer, and a large amount of photovoltaic will be connected. The intermittence, randomness and volatility of photovoltaic power generation can easily lead to problems such as an increase in peak-valley difference, transformer overload and reverse power transmission in the substation.

[0003] Configuring an energy storage system in a photovoltaic substation can better solve the above problems. However, the investment cost of the energy storage system is currently high, and economic problems need to be considered when carrying out capacity configuration. At present, electric vehicles are developing rapidly, and electric vehicles will be widely connected in the substation. The electric vehicle has both load and energy storage system attributes, and is a typical flexible resource. By tapping the adjustable potential of the electric vehicle in the substation, the capacity of the energy storage system can be reduced, the carrying capacity of the substation can be improved, and the reconstruction of the substation can be delayed. Therefore, it is of great practical significance to study a substation energy storage capacity optimization configuration method considering the adjustable potential of the electric vehicle.

[0004] For the capacity optimization configuration of an energy system containing flexible resources, the existing literature is combed, and the following conclusions can be drawn: (1) At present, many scholars have studied the optimization configuration of energy storage capacity, but there is a lack of systematic research on the optimization configuration method of the energy storage capacity of a substation-level energy system. (2) The uncertainty factor processing method based on the probability density function or the membership function has poor feasibility in engineering practice because it depends on the sample size or prior knowledge. (3) Most of the multi-objective optimization algorithms based on population optimization used in the existing literature have the disadvantages of poor convergence and dispersed solution set. SUMMARY

[0005] The technical problem to be solved by the application is to provide a county photovoltaic substation energy storage capacity optimization configuration method containing electric vehicles, which has good convergence and high concentration of solution set.

[0006] To solve the above technical problems, the technical solution adopted by the application is as follows: S1: establishing mathematical models of each flexible resource in the substation and a dispatchable potential model, including a distributed photovoltaic power generation model, a substation load model, an energy storage adjustable potential model, an electric vehicle charging and discharging model and an electric vehicle dispatchable potential model;

[0007] S2: establishing a double-layer optimization configuration model of the energy storage capacity, wherein the upper layer is a capacity optimization configuration model, and the lower layer is an optimization scheduling model;

[0008] S3: The upper model takes the energy storage installed capacity as a decision variable, considers the energy storage installed capacity constraint and other constraint conditions, establishes a peak load shifting objective function and a new energy consumption objective function;

[0009] S4: The lower model takes the electric vehicle charging and discharging power and the energy storage charging and discharging power as decision variables, considers the power balance constraint, the power interaction constraint with the power grid, the energy storage device charging and discharging constraint, the energy storage device power constraint and the electric vehicle charging power constraint, and establishes a minimum substation operation cost objective function;

[0010] S5: The historical records of each electric vehicle are obtained through Monte Carlo random sampling;

[0011] S6: The upper model of the energy storage capacity double-layer optimization configuration model is solved: firstly, the NSGA-II algorithm based on the knee point and the objective weighting method are used to solve the peak load shifting objective function and the new energy consumption objective function and the constraint conditions in step S3; secondly, the particle swarm algorithm is used to solve the minimum substation operation cost and the constraint conditions in step S4;

[0012] S7: The information gap decision method is used to process the uncertainty of photovoltaic and load, so as to determine the uncertainty in the optimization scheduling model.

[0013] Further, the step S1: the distribution function expression of the distributed photovoltaic power generation model is the following formula (11),

[0014]

[0015] Wherein, P v is the photovoltaic output; μ t1 and σ t1 The expression is as follows formula (11D):

[0016]

[0017] In the formula, m t , n t are the mean and variance of the logarithmic normal random variable at t period;

[0018] The substation load model is described by using a normal distribution probability model, and the expression is as follows formula (12),

[0019]

[0020] In formula (12), x t is the load size at t period; μ t2 and σ t2 are the mean and variance of the load at t period, respectively;

[0021] The energy storage adjustable potential model expression is the following formula (13),

[0022]

[0023] In formula (13), e i- , e i+ are upper and lower limits of the energy storage device capacity; p i- , p i+ are upper and lower limits of the energy storage device charge and discharge power; SOC i is the state of charge of the i-th energy storage device; E i is the rated capacity of the i-th energy storage device.

[0024] The electric vehicle charge and discharge model includes a battery charge and discharge model and a battery state of charge model; the battery state of charge model is shown in the following formula (14),

[0025]

[0026] In the formula, is the state of charge of the k-th electric vehicle battery at time t; is the battery capacity of the k-th electric vehicle at time t; C k,0 is the rated capacity of the k-th electric vehicle battery.

[0027] The battery charge and discharge model is divided into a static charge and discharge model and a driving discharge model; the static charge and discharge model is shown in the following formula (15),

[0028]

[0029] In the formula, and are the charge and discharge power of the electric vehicle; η charge and η dis are the charge and discharge efficiencies of the electric vehicle.

[0030] The driving discharge model is shown in the following formula (16),

[0031]

[0032] In the formula, SOC st is the initial state of charge of the electric vehicle when connected to the grid; SOC end is the state of charge of the electric vehicle when disconnected from the grid; L d is the daily driving distance; L0 is the rated distance.

[0033] The electric vehicle dispatchable potential model is shown in the following formula (17),

[0034] Λ v ={P j,t,ch,maxE j,t,min ,E j,t,max ,ΔE j,t} (17)

[0035] P j,t,ch,max is the maximum charging power of the electric vehicle at time t; E j,t,min and E j,t,max are the electric quantity boundaries of the electric vehicle at time t; ΔE j,t is the electric quantity change of the electric vehicle at time t.

[0036] Further, in the distributed photovoltaic power generation model, the expression of photovoltaic output is shown in the following formula (11A),

[0037]

[0038] Q s is the standard solar radiation intensity; r c is the specific solar radiation intensity; W sr is the rated output power of the photovoltaic array; Q t is the solar radiation intensity at time t;

[0039] The distributed photovoltaic measurement and calculation model is shown in the following formula (11B) and (11C),

[0040]

[0041] Q=S[sin(α+β) / sinα]+cosγ*H / tanα (11C)

[0042] In the formula, L is the photovoltaic power generation power, W is the roof photovoltaic installed capacity, Q is the solar radiation intensity, η is the total efficiency of the photovoltaic, S is the horizontal plane solar direct radiation intensity, α is the solar elevation angle, β is the photovoltaic array inclination angle, γ is the geographic azimuth angle, and H is the array height.

[0043] Further, in the electric vehicle dispatchable potential model, the charging and discharging feasible region of a single electric vehicle changes over time, and the power feasible region of a unit device at multiple times is described by the change as shown in the following formula (17C),

[0044]

[0045] In the formula, and are the maximum and minimum electric quantities of the i-th vehicle at time t, respectively, and the minimum electric quantity allowed by the vehicle and the electric quantity when the vehicle owner expects to be off-grid; is the electric quantity when the electric vehicle is connected to the grid; T o,i and T eq,i are the time periods when the electric vehicle is connected to the grid and off-grid, respectively; Pi EV,max P i EV,min P and P and P

[0046] Further, the step S3: the peak load shifting objective function is shown in the following formula (31),

[0047] minf1=max(P load +P ev +P es -P pv )-min(P load +P ev +P es -P pv ) (31)

[0048] P load is the power consumption; P ev is the power consumption; P es is the power consumption; P pv represents the photovoltaic power generation power;

[0049] The new energy consumption objective function is shown in the following formula (32),

[0050]

[0051] P q is the power consumption; P pv is the total photovoltaic power generation power.

[0052] Further, the step S4: the objective function of the minimum substation operation cost is shown in the following formula (41),

[0053]

[0054] P pv,t and P es,t are the photovoltaic output and the energy storage charging and discharging power at t moment respectively; c pv is the photovoltaic operation cost unit price; c es is the energy storage operation cost unit price.

[0055] Further, the power balance constraint can be expressed as shown in the following formula (41A),

[0056] P pv (t)=Pload (t) + P ev (t) + P es (t) + P pv,grid (t) (41A);

[0057] The power exchanged with the grid constraint is represented as shown in the following formula (41B),

[0058] P pv,grid ≥ 0 (41B)

[0059] In the formula, Ppv,grid is the power exchanged with the grid.

[0060] The charge and discharge constraints of the energy storage device are represented as shown in the following formulas (41C) and (41D),

[0061] P es (t) ≤ P BESS,max (41C)

[0062] P dis (t) ≤ P BESS,max (41D)

[0063] In the formula, P BESS,max is the rated maximum charge and discharge power of the energy storage device.

[0064] The power exchanged with the grid constraint is represented as shown in the following formulas (41E) and (41F),

[0065]

[0066] C i + D i ≤ 1 (41F)

[0067] In the formula, SOC i (t) represents the state of charge of the i-th energy storage device at time t; α and β represent the charge and discharge efficiencies of the energy storage device, respectively; E i is the rated capacity of the i-th energy storage device; C i and D i represent the charge and discharge flags of the i-th energy storage device, respectively.

[0068] The electric vehicle charging power constraint can be represented as shown in the following formula (41G),

[0069] 0 ≤ P ev,t ≤ P ev,t,max (41G)

[0070] In the formula, P ev,t,max is the maximum charging power of the electric vehicle at time t.

[0071] Furthermore, in step S6: the NSGA-II algorithm based on the knee point first determines the maximum value A of the economic objective and the maximum value G of the absorptive objective on the Pareto front, and calculates the distance from each individual on the Pareto front to the AG polar line: ax + by + c = 0 as shown in the following formula (61).

[0072]

[0073] Next, determine which individual on the leading edge is furthest from the polar line. This point is the knee region point of the leading edge. Set the knee radius r, and the knee region can be represented as shown in the following formula (62).

[0074]

[0075] In the formula: These represent the normalized vector representations of the knee point and the ideal point, respectively.

[0076] The objective weighting method first standardizes the data and distinguishes between positive and negative indicators. Positive indicators can be expressed as equation (63), and negative indicators can be expressed as equation (64). Then, the correlation coefficient between the targets is calculated, as shown in equation (65).

[0077]

[0078] In the formula: x1 and x2 are the first and second indicators, respectively.

[0079] Furthermore, in step S7: assuming u takes the form of a deterministic model... The economic objective function at that time is F o , believes F c To maximize the impact of uncertainty parameter disturbances, the model for handling photovoltaic and load uncertainties using the information gap decision method is shown in equation (71).

[0080]

[0081] In the formula: λ is the objective function deviation factor;

[0082] The uncertainty in the optimization scheduling model is determined by adjusting the deviation factor λ of the optimization objective and solving for the uncertainty α under this deviation, so that the uncertainty α obtained by the optimization model matches the actual uncertainty.

[0083] Furthermore, step S8 is provided: based on steps S6 and S7, analyze the energy storage capacity optimization configuration results when the electric vehicle penetration rate is at different ratios in industrial and commercial transformer areas and residential transformer areas, and perform scheduling result analysis.

[0084] The beneficial effects generated by the above technical scheme are that the mathematical model of various flexible resources in a county photovoltaic area is established and adjustable potential analysis is performed, a double-layer capacity optimization configuration model is established by an information gap decision theory to reduce the peak-valley difference of the area and improve the photovoltaic consumption rate, and the county photovoltaic area capacity optimization configuration model is solved by using a knee point-based NSGA-II multi-objective optimization algorithm and an objective weighting method; the method has the characteristics of good convergence and high solution set concentration, and is feasible in engineering practice. BRIEF DESCRIPTION OF DRAWINGS

[0085] The application will be further described in detail below with reference to the drawings and specific embodiments.

[0086] Figure 1 is a system configuration diagram of the area according to the application;

[0087] Figure 2 is a basic load diagram of the area according to the application;

[0088] Figure 3 is an adjustable potential diagram of a single electric vehicle according to the application;

[0089] Figure 4 is a knee point and knee area diagram of a concave solution set according to the application;

[0090] Figure 5 is a knee point-based NSGA-II multi-objective optimization algorithm diagram according to the application;

[0091] Figure 6 is a Parato solution set diagram of a capacity configuration scheme according to the application;

[0092] Figure 7 is an iteration convergence result diagram of a lower model according to the application;

[0093] Figure 8 is an average operation cost diagram of different risk strategies according to the application;

[0094] Figure 9 is a price curve diagram according to the application;

[0095] Figure 10 is an interactive power diagram with a power grid under 100% penetration according to the application;

[0096] Figure 11 is an interactive power diagram with a power grid under different models according to the application. DETAILED DESCRIPTION

[0097] Embodiment: The county photovoltaic area energy storage capacity optimization configuration method containing electric vehicles includes the following steps.

[0098] S1: establish mathematical models of each elastic resource in the transformer area and a schedulable potential model, as shown in the following table, including: a distributed photovoltaic power generation model, a transformer area load model, a storage energy schedulable potential model, an electric vehicle charging and discharging model, and an electric vehicle schedulable potential model. The establishment steps of each model are described as follows: Figure 1

[0099] S101, a distributed photovoltaic power generation model is established:

[0100] The photovoltaic output is related to photovoltaic installed capacity, geographic location, photovoltaic panel angle, meteorological factors, and the like. The expression of the photovoltaic output is shown in the following formula (11A),

[0101]

[0102] In the formula, Q s is the standard solar radiation intensity, kW / m 2 ; r c is the specific solar radiation intensity, kW / m 2 ; W sr is the rated output power of the photovoltaic array, kW; Q t is the solar radiation intensity in the t period; P v is the photovoltaic output, i.e., the power generation capacity of the photovoltaic power generation system, kW;

[0103] The distributed photovoltaic calculation model is shown in the following formulas (11B) and (11C),

[0104]

[0105] Q = S [sin (α + β) / sin α] + cos γ * H / tan α (11C)

[0106] In the formula, L is the photovoltaic power, kW; W is the roof photovoltaic installed capacity, kW; Q is the solar radiation intensity, kW / m 2 ; η is the total efficiency of the photovoltaic, %; S is the horizontal plane solar direct radiation intensity, kW / m 2 ; α is the solar elevation angle, i.e., the angle with the horizontal plane, °; β is the photovoltaic array inclination angle, i.e., the angle with the horizontal plane, °, since the photovoltaic panel is usually oriented towards the sky, so this angle β is an obtuse angle; γ is the geographic azimuth angle, °, which is the angle between the true north and the target azimuth line in the clockwise direction; H is the array height, m;

[0107] Due to the influence of random changes in meteorological factors, the photovoltaic power generation output has randomness. The distribution function expression of the distributed photovoltaic power generation model is the following formula (11),

[0108] ​

[0109] where P v is the photovoltaic output, kW; μ t1 and σ t1 are the mean and variance of the lognormal random variable for the t period.

[0110]

[0111] where m t and n t are the mean and variance of the lognormal random variable for the t period.

[0112] S102, a load model of a transformer area is established:

[0113] A certain industrial and commercial transformer area and a residential transformer area are taken as examples to analyze the basic load data and draw the load curve as shown in Figure 2 .

[0114] The load power consumption has uncertainty, and the load model of the transformer area is described by using a normal distribution probability model, and the expression is as follows:

[0115]

[0116] In formula (12), x t is the load size of the t period; μ t2 and σ t2 are the mean and variance of the load of the t period, respectively; and the unit of t is h.

[0117] S103: a model of adjustable potential of energy storage is established:

[0118] Considering the influence of parameters of the energy storage device, the energy storage device is charged and discharged under the control of the aggregator, and the expression of the model of adjustable potential of the energy storage is as follows:

[0119]

[0120] In formula (13), e i- and e i+ are the upper and lower limits of the capacity of the energy storage device, kWh; p i- and p i+ are the upper and lower limits of the charging and discharging power of the energy storage device, kW; SOC i is the state of charge of the i th energy storage device, and the value range is 0-1; and E i is the rated capacity of the i th energy storage device, kWh.

[0121] S104: a charging and discharging model of an electric vehicle is established:

[0122] The electric vehicle charging and discharging model comprises a battery charging and discharging model and a battery state of charge model; the battery state of charge model is shown in the following formula (14),

[0123]

[0124] In the formula, is the state of charge of the kth electric vehicle at time t, and the value range is 0-1; is the battery capacity of the kth electric vehicle at time t, kWh; C k,0 is the rated capacity of the battery of the kth electric vehicle, kWh;

[0125] The battery charging and discharging model is divided into a static charging and discharging model and a driving discharging model; the static charging and discharging model is shown in the following formula (15),

[0126]

[0127] In the formula, and are the charging and discharging power of the electric vehicle, kW; η charge and η dis are the charging and discharging efficiency of the electric vehicle, %;

[0128] The driving discharging model is shown in the following formula (16),

[0129]

[0130] In the formula: SOC st is the initial state of charge of the electric vehicle into the network, and the value range is 0-1; SOC end is the state of charge of the electric vehicle out of the network, and the value range is 0-1; L d is the daily driving mileage, km; L0 is the rated mileage, km.

[0131] S105: Establishing an electric vehicle dispatchable potential model:

[0132] The electric vehicle penetration rate directly reflects the development level of the electric vehicle in the area, and the expression of the electric vehicle penetration rate is shown in the following formula (17A),

[0133]

[0134] In the formula: N EV and N R are the number of electric vehicles and the total number of commercial and industrial area vehicles; then the cluster electric vehicle charging power demand in the area at time t is shown in the following formula (17B),

[0135] P EV (t) = P(t) * N EV(B)

[0136] In the formula: P EV P(t) represents the charging power demand of the cluster, in kW; P(t) represents the charging demand of a single vehicle, in kWh.

[0137] The feasible charging and discharging domain of a single electric vehicle changes over time. The change in the domain characterizes the feasible power domain of the unit device over multiple time periods, as shown in equation (17C).

[0138]

[0139] In the formula: and These are the maximum and minimum battery levels of vehicle i during time period t, the minimum battery level allowed for the vehicle, and the battery level the owner expects to have when the vehicle is disconnected from the grid, in kWh. Electricity consumption when connected to the grid, kWh; T o,i With T eq,i These refer to the time periods when electric vehicles are connected to and disconnected from the grid; P i EV,max With P i EV,min These represent the maximum charging and discharging power of the i-car, in kW; and These are the lower limits of electricity consumption pushed forward from the off-grid period and the lower limits of electricity consumption pushed backward from the grid-connected period, respectively, in kWh; and These are the upper and lower limits of the actual power of vehicle i during time period t, in kW;

[0140] The essence of the single electric vehicle regulation potential model is to project the variable space of individual electric vehicles into a hypercube space while preserving the constraints between variables. Therefore, the electric vehicle ensemble can be compressed into a generalized aggregation model, reducing the model's dimensionality. The hypercube space contains all feasible charging and discharging methods, and the parameters of the generalized aggregation model determine the potential of the electric vehicle cluster as a flexible load storage resource.

[0141] Electric vehicle cluster regulation requires superimposing the decision space of individuals as a whole. The Minkowski addition method is applicable to Euclidean space, and its physical essence is an expanded set of multiple spaces. Therefore, it can effectively aggregate electric vehicle groups into generalized devices, and the adjustable potential model of electric vehicle clusters is shown in the following equation (17).

[0142] Λ v ={P j,t,ch,max E j,t,min E j,t,max ,ΔE j,t} (17)

[0143] In the formula: P j,t,ch,maxThe maximum charging power of an electric vehicle during time period t, in kW; E j,t,min and E j,t,max Let ΔE be the boundary of the electric vehicle's charge in time period t, in kWh; j,t Let represent the change in the electric vehicle's charge level during time period t, in kWh.

[0144] Assume the electric vehicle's state of charge (SOC) is when it is connected to the power grid. st Users expect electric vehicles to have a state of charge (SOC) when disconnected from the grid. end The typical charge / discharge boundary model for a single electric vehicle is as follows: Figure 3 As shown. The upper bound of the energy trajectory is: after the electric vehicle enters the grid at time tin, it immediately charges at the rated power until the desired energy level is reached, and then remains unchanged until it leaves the grid at time tout; the lower bound of the energy trajectory is: after the electric vehicle enters the grid at time tin, it discharges and finally charges at the rated power to the desired energy level before leaving the grid; the slope k represents the charging and discharging power value of the electric vehicle, and all regions between the upper and lower bounds of the energy trajectory can be considered as feasible energy trajectories.

[0145] S2: Establish a two-layer optimization configuration model for energy storage capacity, where the upper layer is the capacity optimization configuration model and the lower layer is the optimization scheduling model.

[0146] The energy storage capacity optimization configuration model adopts a two-layer structure. The upper layer is the capacity optimization configuration model, which takes the number of electric vehicles and photovoltaic installed capacity under different penetration rates and charging / discharging modes as inputs, and the installed capacity of energy storage equipment as the decision variable. The optimization objectives are to reduce peak-valley differences and improve photovoltaic absorption rate. The calculated capacity configuration result serves as the input to the lower-layer optimization scheduling model. The lower layer is the optimization scheduling model, which considers the uncertainties of photovoltaic output and load. It takes the charging / discharging power of electric vehicles and the charging / discharging power of energy storage as decision variables, and aims to minimize operating costs. The optimized operating cost is fed back to the upper-layer capacity optimization configuration model. Finally, after satisfying the required number of iterations, the capacity configuration result is output. Considering the influence of uncertainties, the rated power of energy storage is taken as 1.05 to 1.3 times the maximum value of typical daily scheduling instructions, with a best value of 1.1 times.

[0147] S3: The upper-level model uses energy storage installed capacity as the decision variable, considers constraints such as energy storage installed capacity, and establishes peak shaving and valley filling objective functions and new energy consumption objective functions; including the following steps:

[0148] S301, the upper-level model uses energy storage installed capacity as the decision variable:

[0149] A county-level photovoltaic (PV) distribution area capacity optimization configuration model incorporating flexible resources can determine the energy storage capacity based on the penetration rate of electric vehicles and the PV installed capacity. The decision variables of the optimization model can be expressed as shown in equation (30A).

[0150] X = [N es ](30A)

[0151] In the formula: N es is the energy storage installed capacity, unit kWh.

[0152] S302, considering the energy storage installed capacity constraints and other constraints:

[0153] The energy storage capacity configuration model needs to meet certain energy storage installed capacity constraints, and the constraint formula is shown in the following formula (30B),

[0154] N es,min ≤N ess ≤N es,max (30B)

[0155] In the formula: N es,min is the lower limit of energy storage installed capacity, kWh; N es,max is the upper limit of energy storage installed capacity, kWh; N ess is the energy storage capacity, kWh.

[0156] S303, establish the multi-objective function of peak load shifting and new energy consumption:

[0157] 1) Peak load shifting objective function:

[0158] The peak-valley difference is the difference between the electricity load, electric vehicle load, energy storage load and photovoltaic power generation power, and the minimum peak-valley difference is taken as the target, and the peak load shifting objective function is shown in the following formula (31),

[0159] minf1 = max(P load +P ev +P es -P pv )-min(P load +P ev +P es -P pv ) (31)

[0160] In the formula: P load is the electricity load, kW; P ev is the electric vehicle charging and discharging power, kW; P es is the energy storage charging and discharging power, kW; P pv represents the photovoltaic power generation power, kW.

[0161] 2) New energy consumption objective function:

[0162] The photovoltaic consumption rate is the percentage of the total power consumption in the area to the total photovoltaic power generation power, and the maximum photovoltaic consumption rate is taken as the target, that is, the minimum anti- sending power rate, and the new energy consumption objective function is shown in the following formula (32),

[0163]

[0164] In the formula: P q Power consumed within the transformer area, kW; P pv The total photovoltaic power generation capacity is expressed in kW.

[0165] S4: The lower-level model uses the charging and discharging power of electric vehicles and the charging and discharging power of energy storage as decision variables. It considers constraints such as power balance constraints, power interaction constraints with the grid, charging and discharging constraints of energy storage devices, energy storage device capacity constraints, and electric vehicle charging power constraints to establish the objective function for minimizing the operating cost of the distribution area. This includes the following steps:

[0166] S401, the lower-level model uses the charging and discharging power of electric vehicles and the charging and discharging power of energy storage as decision variables:

[0167] The optimized scheduling model, with electric vehicle charging and discharging power and energy storage charging and discharging power as decision variables, can be expressed as shown in equation (40).

[0168] X = [P] ev ,P es (40)

[0169] In the formula: P ev Electric vehicle charging and discharging power, unit: kW; P es This refers to the energy storage charging and discharging power, measured in kW.

[0170] S402 considers constraints such as power balance constraints, power interaction constraints with the grid, energy storage device charging and discharging constraints, energy storage device power constraints, and electric vehicle charging power constraints.

[0171] 1) The power balance constraint can be expressed as shown in the following equation (41A).

[0172] P pv (t)=P load (t)+P ev (t)+P es (t)+P pv,grid (t) (41A);

[0173] 2) The power constraint interacting with the power grid is expressed as shown in the following equation (41B).

[0174] P pv,grid ≥0 (41B)

[0175] In the formula: Ppv,grid is the power interacting with the power grid, in kW;

[0176] 3) The charge and discharge constraints of the energy storage device are expressed as shown in equations (41C) and (41D).

[0177] Pes (t)≤P BESS,max (41C)

[0178] P dis (t)≤P BESS,max (41D)

[0179] P BESS,max is the rated maximum charge-discharge power of the energy storage device, kW;

[0180] 4) The energy storage device power constraint is represented as shown in the following formula (41E) and (41F),

[0181]

[0182] C i +D i ≤1 (41F)

[0183] SOC i (t) represents the charge of the i-th energy storage device at time t, kWh; α and β represent the charge-discharge efficiency of the energy storage device, %; E i is the rated capacity of the i-th energy storage device, kWh; C i and D i represent the charge-discharge flag of the i-th energy storage device, respectively.

[0184] 5) The electric vehicle charging power constraint can be represented as shown in the following formula (41G),

[0185] 0≤P ev,t ≤P ev,t,max (41G)

[0186] P ev,t,max is the maximum charging power of the electric vehicle at time t, kW.

[0187] S402, a target function of minimum substation operation cost is established:

[0188] Taking the minimum substation operation cost as the optimization target, the target function of the minimum substation operation cost is shown in the following formula (41),

[0189]

[0190] P pv,t and P es,t are the photovoltaic output and the energy storage charge-discharge power at time t, kW; c pv is the photovoltaic operation cost unit price, yuan / kWh; c es is the energy storage operation cost unit price, yuan / kWh.

[0191] S5: Obtain the historical records of each electric vehicle through Monte Carlo random sampling, including: the number of electric vehicles Nev, the daily driving mileage of electric vehicles L, the time of electric vehicles entering the grid, the time of electric vehicles leaving the grid, the electric quantity of electric vehicles when accessing the grid, and the electric quantity of electric vehicles when the owner expects to leave the grid; obtain the basic parameter data of the unit capacity cost of energy storage, the construction unit price of photovoltaic, the charging and discharging efficiency of energy storage, and the SOC range of energy storage.

[0192] S6: Solve the upper model of the energy storage capacity bi-level optimization configuration model: first, according to the peak clipping and valley filling objective function and the new energy consumption objective function and the constraint conditions in step S3, use the NSGA-II algorithm based on the knee point and the objective weighting method to solve; second, according to the minimum operating cost of the transformer station and the constraint conditions in step S4, use the particle swarm algorithm to solve; including the following steps:

[0193] S601, NSGA-II algorithm based on knee point:

[0194] Embedding the knee point selection mechanism into the NSGA-II multi-objective optimization algorithm can make the solution set more centralized and the convergence better, and the performance change after normalizing the objective function is not biased to any objective, which can balance the conflict between objectives. First, determine the maximum value A of the economic objective and the maximum value G of the consumption objective on the Pareto frontier, and calculate the distance of each individual on the Pareto frontier to the AG polar line: ax+by+c=0 as shown in the following formula (61),

[0195]

[0196] Then determine which individual on the frontier has the farthest distance to the polar line, and the point is the knee point of the frontier. Set the knee radius r, then the knee area can be represented as shown in the following formula (62),

[0197]

[0198] In the formula: respectively represent the normalized vector representation of the knee point and the ideal point. Figure 4 The knee point and the knee area under the Pareto concave solution set, this mechanism can accelerate the convergence of the algorithm and improve the efficiency of the algorithm; the algorithm flow is as shown in Figure 5 .

[0199] S602, objective weighting method:

[0200] Criteria Importance Though Intercrieria Correlation (CRITIC) is used to solve the problem of multi-index comprehensive evaluation of the size of each index weight, which can help to select the optimal scheme. First, the data is standardized, and the positive and negative of the index is distinguished. The larger the positive index is, the better it is, and the smaller the negative index is, the better it is, as shown in the following formula (63) and formula (64); then the correlation coefficient between the targets is calculated, as shown in the following formula (65). This embodiment has one object to be evaluated and two evaluation indexes, which can form a data matrix X=(x ij )1×2, and the elements in the data matrix are x' ij .

[0201] 1) The positive index can be represented by the following formula (63),

[0202]

[0203] 2) The negative index can be represented by the following formula (64),

[0204]

[0205] 3) The correlation coefficient between the targets is shown in the following formula (65),

[0206]

[0207] In the formula, x1 and x2 are the first index and the second index respectively; the first index is the peak clipping and valley filling target, and the second index is the new energy consumption target.

[0208] S7: The information gap decision method is used to process the uncertainty of photovoltaic and load, so as to determine the uncertainty in the optimization scheduling model. Including the following steps:

[0209] S701, the information gap decision method is used to process the uncertainty of photovoltaic and load,

[0210] The information gap decision method (IGDT) can well depict the uncertainty of the probability distribution function which is difficult to obtain, and can maximize the change of uncertainty while ensuring the expected target of the model. The optimization model based on IGDT can be represented by the following formula (71A),

[0211]

[0212] In the formula, F is the objective function; X is the decision variable; u is the uncertainty matrix; h and g are model constraints.

[0213] The uncertainty of IGDT is generally described by non-probabilistic models, including envelope model, fractional uncertainty model, sphere model, etc. In this embodiment, the fractional uncertainty model is adopted, and the fluctuation of the actual value of the uncertainty quantity u around its predicted value may be expressed as shown in the following formula (71B),

[0214]

[0215] In the formula, a is the uncertainty. According to the historical data of the photovoltaic output and the power load of a certain area, the uncertainty of the photovoltaic output is 0.059 and the uncertainty of the power load of the area is 0.032 at a 95% confidence level.

[0216] Suppose that the economic objective function of u taking in the deterministic model is F o , and F c is considered as the maximum of the uncertainty parameter disturbance, then the model for processing the uncertainty of photovoltaic and load by using the information gap decision method is shown in the following formula (71),

[0217]

[0218] In the formula, λ is the deviation factor of the objective function, that is, the deviation degree between the expected target and the optimal target of the deterministic model.

[0219] S702: Determine the uncertainty in the optimization scheduling model: by adjusting the deviation factor λ of the optimization target, the uncertainty α under the deviation is solved, so that the uncertainty α obtained by the optimization model is consistent with the actual situation.

[0220] The uncertainty of photovoltaic and load at a 95% confidence level is determined by using the information gap decision method as [0.059, 0.032], and the uncertainty weight of photovoltaic and load is determined by using the objective weighting method as [0.675, 0.325], so that the uncertainty in the optimization scheduling model is α = 0.059 * 0.675 + 0.032 * 0.325 = 0.05. By adjusting the deviation factor λ of the optimization target, the uncertainty α under the deviation is solved, so that the uncertainty α obtained by the optimization model is consistent with the actual situation, wherein the uncertainty α and the deviation factor λ are positively correlated. The deviation factor of the risk-averse strategy is 0.05 and the average operating cost is 0.0457 yuan / kWh when α = 0.05 is solved by using the interpolation method.

[0221] S8: According to steps S6 and S7, analyze the energy storage capacity optimization configuration results of the electric vehicle penetration rate at different proportions in the industrial and commercial area and the residential area, preferably the energy storage capacity optimization configuration results of the electric vehicle penetration rate of 20%, 50%, and 100%, and perform scheduling result analysis. Including the following steps:

[0222] S801: An example study is carried out for a certain county photovoltaic area, the geographical position of which is east longitude 113°43', north latitude 34°26', transformer capacity 2000kVA, load rate not more than 90%, according to the maximum planning area of distributed energy storage, the optimization interval of distributed energy storage installed capacity Z is calculated as [0kWh, 1000kWh]; wherein, the model parameter settings are shown in Table 1;

[0223] Table 1: Model parameter settings

[0224]

[0225]

[0226] In order to measure the difference of electric vehicle charging mode, in addition to the objective function, three indexes (investment cost, CO2 emission and daytime reverse power sending period) are selected to evaluate and compare the energy storage capacity configuration results. Among them, the investment cost of photovoltaic can be expressed as shown in the following formula (81),

[0227]

[0228] In the formula: C pv1 is the construction unit price of photovoltaic power generation, yuan / kW; C pv2 is the replacement cost of photovoltaic power generation, yuan / kW·year; N pv is the quotient of construction unit price and life cycle, yuan / year; T pv is the life cycle of photovoltaic power station, years; r is the discount rate, %.

[0229] S802, double-layer model solving:

[0230] The upper model is solved by using the NSGA-II algorithm based on the knee point, and the NSGA-II algorithm program is run in the MATLAB environment, and the algorithm parameter settings are shown in Table 2;

[0231] Table 2: NSGA-II algorithm parameter settings

[0232] Parameter Maximum iteration number Population size Crossover operator Mutation probability Credibility threshold Value 200 100 2 0.8 0.8

[0233] Assuming that the electric vehicle penetration rate in the area is 20%, 50% and 100%, the number of electric vehicles is 20, 50 and 100 respectively. For the current resource composition of the area, taking the scenario of electric vehicle penetration rate of 100% in industrial and commercial area and charging and discharging behavior as an example, the Parato solution set of the capacity configuration scheme is as follows Figure 6The objective weight is w = {0.6854, 0.3146} obtained by formula (63)-(65), and the optimal scheme is obtained by multiplying and accumulating the objective weight of each candidate solution in the Parato solution set with the objective result and sorting.

[0234] The lower model is solved by using the particle swarm algorithm, and the particle swarm algorithm program is run in the MATLAB environment, and the algorithm parameters are set as shown in Table 3;

[0235] Table 3: Particle swarm algorithm parameter setting

[0236] Parameter Maximum iteration number Population size Learning factor Inertia weight Dimension D Value 100 200 2 0.8 24

[0237] The iteration convergence result obtained by the lower model with the minimum power grid operation cost as the target is shown in Table 2, and the objective function value decreases with the increase of the iteration number and gradually tends to be stable. Figure 7

[0238] S803, strategy parameter processing:

[0239] The information gap decision method is used to determine the uncertainty of photovoltaic and load under 95% confidence level, which is [0.059, 0.032], and the objective weighting method is used to determine the uncertainty weight of photovoltaic and load, which is [0.675, 0.325], so as to determine the uncertainty α in the optimization scheduling model, which is α = 0.059*0.675 + 0.032*0.325 = 0.05. By adjusting the deviation factor λ of the optimization objective, the uncertainty α under the deviation is solved, so that the uncertainty α obtained by the optimization model is consistent with the actual situation; Figure 8 Taking the operation result of electric vehicles participating in charging and discharging as an example, the uncertainty and average operation cost of the system under the change of the deviation factor of different strategies are obtained, wherein the uncertainty α is positively correlated with the deviation factor λ. By using the interpolation method, the deviation factor of the risk aversion strategy is 0.05 when α = 0.05, and the average operation cost is 0.0457 yuan / kWh.

[0240] S804, capacity configuration result analysis:

[0241] Since electric vehicles have the behaviors of unordered charging, ordered charging and ordered charging and discharging, the energy storage capacity optimization configuration results of electric vehicles with penetration rates of 20%, 50% and 100% in industrial and commercial areas and residential areas are shown in Tables 4 and 5, respectively, and the unit is kW / kWh.

[0242] Table 4: Capacity configuration results of industrial and commercial areas under different penetration rates (unit: kW / kWh)

[0243] ​

[0244] From Table 4, in the case of unordered charging of electric vehicles, the storage capacity configuration under 20%, 50% and 100% penetration rates of industrial and commercial areas is 812.02 kWh, 867.04 kWh and 997.52 kWh, respectively, and the storage power is 397.48 kW, 431.30 kW and 491.76 kW, respectively. Due to the increase in charging power of electric vehicles caused by the increase in penetration rate, the installed capacity of storage is also increased. For orderly scheduling of electric vehicles, excess photovoltaic power during periods of high photovoltaic output can be effectively absorbed, and the reduced storage power and capacity under different penetration rates are 34.49 kW / 46.39 kWh, 89.07 kW / 46.39 kWh and 187.32 kW / 46.39 kWh, respectively. The orderly charging and discharging mode has obvious advantages in storage capacity configuration, and compared with orderly charging, the orderly charging and discharging mode reduces the storage power and capacity by 44.08 kW / 52.69 kWh, 58.07 kW / 93.68 kWh and 92.24 kW / 184.63 kWh, respectively, under three penetration rates.

[0245] Table 5: Capacity configuration results of residential areas under different penetration rates (unit: kW / kWh)

[0246]

[0247] As can be seen from Table 5, due to the influence of regional functionality and grid-connected / off-grid time on the charging and discharging of electric vehicles in residential areas, the storage capacity configuration under unordered charging of electric vehicles is higher than that in industrial and commercial areas. The reduced storage power and capacity under orderly charging of electric vehicles compared to unordered charging under different penetration rates are 45.65 kW / 54.67 kWh, 87.17 kW / 129.64 kWh and 174.90 kW / 295.55 kWh, respectively. The orderly charging and discharging mode of electric vehicles not only plays a "filling valley" role in absorbing excess photovoltaic power through charging behavior when the basic load is not high at noon, but also plays a "peak shaving" role through discharging behavior when the load is high at night. Compared with orderly charging, the orderly charging and discharging mode reduces the storage power and capacity by 65.68 kW / 105.93 kWh, 89.74 kW / 150.55 kWh and 103.45 kW / 206.42 kWh, respectively, under three penetration rates, and the mode can further effectively reduce the installed capacity of storage through discharging behavior. Electric vehicles and distributed storage can be coordinated and scheduled to reduce electricity costs during peak periods, while better utilizing renewable energy during off-peak periods.

[0248] S805, analyze the scheduling result.

[0249] With the data of 20% and 100% penetration rate of electric vehicles in industrial and commercial area as input, the configuration results under two penetration rates are optimized and scheduled under different charging modes, and the specific scheduling results are shown in Table 6 below;

[0250] Table 6: Optimization scheduling results

[0251]

[0252] As shown in Table 6, from the economic analysis of operation, the investment cost of orderly charging under 20% and 100% penetration rate is reduced by 51,000 yuan and 395,000 yuan respectively compared with unordered charging; the investment cost of orderly charging and discharging is reduced by 58,000 yuan and 20,3,000 yuan respectively compared with orderly charging. The orderly scheduling and charging and discharging behavior of electric vehicles can effectively reduce the capacity configuration of energy storage, and reduce the investment cost of the area.

[0253] From the load analysis, the peak-valley difference of orderly charging under two penetration rates is reduced by 226.7 kW and 552.7 kW respectively compared with unordered charging; the peak-valley difference of orderly charging and discharging is reduced by 117.8 kW and 273.1 kW respectively compared with orderly charging. In the orderly charging mode, electric vehicles can be fully scheduled to charge during the low load period of the area, and the charging behavior is reduced during the peak load period, thereby reducing the peak-valley difference of the area. In the orderly charging and discharging mode, electric vehicles can be scheduled to discharge during the peak load period, further reducing the peak-valley difference and playing the role of peak shaving and valley filling.

[0254] From the perspective of consumption, in the unordered charging mode of electric vehicles, the situation of anti-power sending caused by difficulty in photovoltaic consumption occurs, which affects the safe and stable operation. The optimization scheduling of electric vehicles can guide more electric vehicles to charge under the condition of large photovoltaic power generation at noon, thereby consuming excess photovoltaic and reducing the anti-power sending situation, thereby improving the stability and environmental friendliness of the system.

[0255] From Figure 9 and Figure 10 It can be seen that most of the electric vehicles charge at the time of valley electricity price and discharge at the time of peak electricity price, so that load shifting can be realized. Compared with unordered charging and orderly charging mode, the interactive power with the grid under the multi-objective optimization of the orderly charging and discharging mode of electric vehicles changes. After scheduling, the overall area realizes peak shaving and valley filling, and there is no anti-power sending situation. For the scheduling after capacity optimization configuration, the charging and discharging behavior of electric vehicles is more in line with the demand of the area. By fully tapping the adjustable potential of flexible resources, the peak-valley difference of the area can be effectively reduced, the photovoltaic consumption rate can be improved, and the operation cost can be reduced, which proves the effectiveness of the proposed method.

[0256] S806, Deterministic vs. IGDT

[0257] To investigate the role of IGDT model in this embodiment, the energy storage configuration results are compared between the deterministic model and the IGDT model. For the case of 100% electric vehicle penetration rate and orderly charging and discharging in industrial and commercial areas, the average operating cost and energy storage capacity configuration results of the deterministic model and the IGDT model are listed in Table 7.

[0258] Table 7: Planning results of different models

[0259] Model Average operating cost (yuan / kWh) Energy storage capacity configuration (kWh) Deterministic 0.0422 458.38 IGDT 0.0457 503.95

[0260] From Table 7, it can be seen that because the IGDT model considers the uncertainty of photovoltaic output and load, the energy storage capacity configuration is increased compared with the deterministic model, and accordingly, the average operating cost is also increased. For the two capacity configuration results, optimization scheduling is carried out in the extreme scenario, i.e. small photovoltaic output and large load demand, and the obtained grid interaction power is shown in Table 8. Figure 11 It can be seen that in the extreme scenario operation mode, the photovoltaic output is reduced, and the photovoltaic consumption rate is very high. The peak-valley difference of the IGDT model is reduced by 138.13 kW compared with the deterministic model. This shows that because the IGDT model has more energy storage capacity configuration, the obtained scheduling result has better adaptability and better robustness for the operation mode in the extreme scenario.

Claims

1. A method for optimizing the energy storage capacity configuration of a county-level photovoltaic power station area including electric vehicles, characterized in that, It includes the following steps: S1: Establish mathematical models and dispatchable potential models for each flexible resource within the distribution area, including: distributed photovoltaic power generation model, distribution area load model, energy storage adjustable potential model, electric vehicle charging and discharging model, and electric vehicle dispatchable potential model; S2: Establish a two-layer optimization configuration model for energy storage capacity, where the upper layer is the capacity optimization configuration model and the lower layer is the optimization scheduling model; S3: The upper-level model uses energy storage capacity as the decision variable, considers energy storage capacity constraints, and establishes peak shaving and valley filling objective functions and photovoltaic consumption objective functions. The objective function for peak shaving and valley filling is shown in equation (31). minf1=max(P load +P ev +P es -P pv )-min(P load +P ev +P es -P pv ) (31) In the formula: P load For electrical load; P ev Power for charging and discharging electric vehicles; P es For energy storage charging and discharging power; P pv Indicates photovoltaic power generation capacity; The photovoltaic power consumption objective function is shown in equation (32). In the formula: P q Power consumed within the transformer area; P pv This represents the total power generated by photovoltaic power generation. S4: The lower-level model uses the charging and discharging power of electric vehicles and the charging and discharging power of energy storage as decision variables. It considers the constraints of power balance, power interaction with the grid, charging and discharging constraints of energy storage devices, power constraints of energy storage devices, and charging power constraints of electric vehicles to establish the objective function of minimizing the operating cost of the transformer substation. The objective function for minimizing the operating cost of a transformer substation is shown in equation (41). In the formula: P pv,t and P es,t These represent the photovoltaic power output and energy storage charging / discharging power at time t, respectively; c pv c is the unit price of photovoltaic operating costs; es This refers to the unit price of energy storage operation costs; S5: Obtain the historical records of each electric vehicle through Monte Carlo random sampling; S6: Solve the upper-level model of the dual-layer optimization configuration model for energy storage capacity: First, based on the peak shaving and valley filling objective function, the photovoltaic absorption objective function, and the constraints in step S3, use the NSGA-II algorithm based on the knee point and the objective weighting method to solve the problem; second, based on the minimum operating cost of the transformer area and the constraints in step S4, use the particle swarm optimization algorithm to solve the problem. S7: The information gap decision method is used to handle the uncertainties of photovoltaics and loads, thereby determining the uncertainties in the optimal scheduling model.

2. The method for optimizing the energy storage capacity of a county-level photovoltaic power station area including electric vehicles, as described in claim 1, is characterized in that, Step S1: The distribution function expression of the distributed photovoltaic power generation model is as follows (11), Among them, P v Powering photovoltaics; μ t1 and σ t1 The expression is as follows (11D): Where: m t n t Let be the mean and variance of a log-normal random variable over time period t; The load model of the transformer area is described by a normal distribution probability model, and the expression is as follows (12). In equation (12): x t The load size during time period t; μ t2 and σ t2 These are the mean and variance of the load during time period t, respectively. The expression for the adjustable energy storage potential model is as follows (13): In equation (13): e i+、 e i- These represent the upper and lower limits of the energy storage device's capacity, respectively; p i+、 p i- These represent the upper and lower limits of the charging and discharging power of the energy storage device; SOC i E represents the state of charge of the i-th energy storage device; i e represents the rated capacity of the i-th energy storage device; i For energy storage device capacity; p i The charging and discharging power of energy storage devices; The electric vehicle charging and discharging model includes a battery charging and discharging model and a battery state of charge model; the battery state of charge model is shown in equation (14) below. In the formula: Let t be the state of charge of the battery of the kth electric vehicle. Let be the battery charge of the k-th electric vehicle at time t; The rated capacity of the battery of the kth electric vehicle; The battery charging and discharging model is divided into a static charging and discharging model and a driving discharging model; the static charging and discharging model is shown in the following formula (15). In the formula: and These represent the charging and discharging power of the electric vehicle; η charge and η dis These are the charging and discharging efficiencies of electric vehicles, respectively. The driving discharge model is shown in equation (16). Where: SOC st The initial state of charge (SOC) of an electric vehicle upon grid connection. end This refers to the off-grid state of charge of electric vehicles; L d L0 represents the daily mileage; L0 represents the rated mileage. The schedulable potential model for electric vehicles is shown in equation (17). Λ v ={P j,t,ch,max ,E j,t,min ,E j,t,max ,ΔE j,t } (17) In the formula: P j,t,ch,max E represents the maximum charging power of an electric vehicle during time period t. j,t,min and E j,t,max Let ΔE be the boundary of the electric vehicle's charge level during time period t; j,t Let represent the change in the electric vehicle's charge level during time period t.

3. The method for optimizing the energy storage capacity of a county-level photovoltaic power station area including electric vehicles, as described in claim 2, is characterized in that... In the distributed photovoltaic power generation model, the expression for photovoltaic output is shown in equation (11A). In the formula: Q s The solar radiation intensity under standard conditions; r c For a specific solar radiation intensity; W sr Q represents the rated output power of the photovoltaic array. t The solar radiation intensity during time period t; The calculation model for distributed photovoltaic power generation is shown in equations (11B) and (11C). Q=S[sin(α+β) / sinα]+cosγ*H / tanα(11C) In the formula: L is the photovoltaic power generation capacity; W is the rooftop photovoltaic installed capacity; Q is the solar radiation intensity; η is the total photovoltaic efficiency; S is the horizontal direct solar radiation intensity; α is the solar altitude angle; β is the photovoltaic array tilt angle; γ is the geographical azimuth angle; H is the array height.

4. The method for optimizing the energy storage capacity of a county-level photovoltaic power station area including electric vehicles according to claim 1, characterized in that: The power balance constraint is expressed as shown in equation (41A). P pv (t)=P load (t)+P ev (t)+P es (t)+P pv,grid (t) (41A); The power constraint for interaction with the power grid is expressed as shown in equation (41B). P pv,grid ≥0(41B) In the formula: Ppv,grid represents the power interacting with the power grid; The charge and discharge constraints of the energy storage device are expressed as shown in the following formula (41C). P es (t)≤P BESS,max (41C) In the formula: P BESS,max The rated maximum charge and discharge power of the energy storage device; The energy storage device's power constraint is expressed as shown in equations (41D) and (41E). C i +D i ≤1(41E) Where: SOC i (t) represents the charge of the i-th energy storage device at time t; α and β represent the charging and discharging efficiency of the energy storage device, respectively; E i C represents the rated capacity of the i-th energy storage device; i D i These represent the charging and discharging flags of the i-th energy storage device; The electric vehicle charging power constraint is expressed as shown in the following formula (41G). 0≤P ev,t ≤P ev,t,max (41G) In the formula: P ev,t,max The maximum charging power for an electric vehicle at time t.

5. The method for optimizing the energy storage capacity of a county-level photovoltaic power station area including electric vehicles according to claim 1, characterized in that, Step S6: The NSGA-II algorithm based on the knee point first determines the maximum value A of the economic objective and the maximum value G of the absorptive objective on the Pareto front. The distance from each individual on the Pareto front to the AG polar line: ax + by + c = 0 is calculated as shown in the following formula (61). Next, determine which individual on the leading edge is furthest from the polar line. This point is the knee region point of the leading edge. Set the knee radius r, and the knee region is represented as shown in the following formula (62). In the formula: These represent the normalized vector representations of the knee point and the ideal point, respectively. The objective weighting method first standardizes the data and distinguishes between positive and negative indicators. Positive indicators are represented by equation (63), and negative indicators are represented by equation (64). Then, the correlation coefficient between the targets is calculated, as shown in equation (65). In the formula: x1 and x2 are the first and second indicators, respectively.

6. The method for optimizing the energy storage capacity of a county-level photovoltaic power station area including electric vehicles according to claim 1, characterized in that, Step S7: Assuming a deterministic model, u takes... The economic objective function at that time is F o , believe F c To maximize the impact of uncertainty parameter disturbances, the model for handling photovoltaic and load uncertainties using the information gap decision method is shown in equation (71). In the formula: λ is the objective function deviation factor; The uncertainty in the optimization scheduling model is determined by adjusting the deviation factor λ of the optimization objective and solving for the uncertainty α under this deviation, so that the uncertainty α obtained by the optimization scheduling model matches the actual uncertainty.

7. A method for optimizing the energy storage capacity of a county-level photovoltaic power station area containing electric vehicles, as described in any one of claims 1-6, characterized in that: Step S8 is also included: Based on steps S6 and S7, analyze the energy storage capacity optimization configuration results when the electric vehicle penetration rate is at different ratios in industrial and commercial transformer areas and residential transformer areas, and perform scheduling result analysis.

Images