Residential area-working area interactive electric vehicle dynamic pricing method for promoting peak load shifting and distributed photovoltaic consumption

Abstract

The invention discloses a residential area-working area interactive electric vehicle dynamic pricing method for promoting peak load shifting and distributed photovoltaic consumption. The method comprises the steps of formulating a dynamic time-of-use electricity price for an electric vehicle user parked in a working area for charging; fixed time-of-use electricity prices are formulated for the electric vehicle users selected to be charged in the residential area; establishing an electric vehicle user demand-price response model under regional electricity prices to measure the sensitive degree of the electricity consumption of the electric vehicle user to the change of the electricity price; and with the maximum photovoltaic utilization rate of the working area, the minimum peak-valley difference of the equivalent load of the residential area and the minimum charging cost of the electric vehicle user as targets, a multi-objective function is established to perform optimal scheduling on the electric vehicle. And a multi-objective optimization problem is solved by adopting NSGA-II with an elitist strategy. According to the method, on the basis of the condition that an electric vehicle owner tends to charge in a period of low electricity price, the charging place and time of the electric vehicle are transferred through a mechanism of formulating the electricity price in different regions, and optimal distribution of the electric vehicle in the time dimension and the space dimension is achieved.

Description

Technical Field

[0001] The present invention relates to the field of coordinated renewable energy consumption by electric vehicles, and in particular to a residential-workspace interactive electric vehicle dynamic pricing method that promotes peak shaving and valley filling and distributed photovoltaic consumption. Background Art

[0002] Electric vehicles have gained widespread popularity in recent years as a green alternative to traditional fossil fuel vehicles and a viable solution for reducing urban air pollution. With the rapid growth in electric vehicle ownership, demand for charging has increased, primarily in residential areas during the evening peak hours. The combined charging load of electric vehicles and conventional loads creates a "peak upon peak" phenomenon during peak electricity demand, leading to excessive loads during peak hours and impacting the safety and stability of residential power distribution facilities.

[0003] Reference [1]: Feng Congcong. Adjustment of time-of-use electricity price policy moves towards refinement [N]. China Electric Power News, 2025-05-09(002). It points out that with the rapid growth of photovoltaic installed capacity, the power surplus during the peak photovoltaic power generation period at noon has become more and more obvious, and the problem of temporal and spatial mismatch between power supply and demand has become more and more prominent. In the future smart grid, work areas will become the first choice for using distributed photovoltaic power generation to charge electric vehicles. The mismatch between electric vehicle electricity consumption and photovoltaic output will further aggravate the temporal and spatial mismatch between power supply and demand. It is difficult to adapt to the sharp fluctuations on both the source and load sides by relying solely on the traditional time-of-use electricity price mechanism.

[0004] It can be seen from the travel characteristics of electric vehicle users that a large number of electric vehicle owners will arrive and park in the work area during the day. Therefore, it is possible to attract more electric vehicle users to charge in the work area by guiding regional electricity prices, thereby promoting distributed photovoltaic consumption while avoiding a large number of car owners gathering in residential areas to charge. Summary of the Invention

[0005] To address the problem of mismatch between electric vehicle load and distributed photovoltaic output in terms of time and space, the present invention proposes a residential-workspace interactive electric vehicle dynamic pricing method that promotes peak shaving and valley filling and distributed photovoltaic consumption. This method is based on the fact that electric vehicle owners tend to charge during periods of lower electricity prices. By setting electricity prices in different regions, the charging location and time of electric vehicles are shifted, thereby achieving optimal allocation of electric vehicles in both time and space dimensions.

[0006] The technical solution adopted by the present invention is:

[0007] A residential-workspace interactive electric vehicle dynamic pricing method for promoting peak shaving and valley filling and distributed photovoltaic consumption includes the following steps:

[0008] Step 1: For electric vehicle users charging at work areas, a dynamic time-of-use electricity price is set based on the day-ahead forecast of PV output. For electric vehicle users charging in residential areas, a fixed time-of-use electricity price is set based on the day-ahead forecast of the residential area's basic electricity load.

[0009] Step 2: Establish an electric vehicle user demand-price response model under regional electricity prices to measure the sensitivity of electric vehicle users' electricity consumption to electricity price changes;

[0010] Step 3: With the goals of maximizing the photovoltaic utilization rate in the work area, minimizing the peak-to-valley difference of the equivalent load in the residential area, and minimizing the charging cost for electric vehicle users, a multi-objective function is established to optimize the scheduling of electric vehicles.

[0011] Step 4: Use the non-dominated sorting genetic algorithm with elite strategy (NSGA-Ⅱ) to solve the multi-objective optimization problem. In step 1, in order to characterize the relationship between photovoltaic output and electricity price, the basic charging electricity price is first associated with photovoltaic output, and the initial electricity price c for each period of the working area is defined. 0,t for:

[0012]

[0013] In formula (1), c0 is the basic charging electricity price, that is, the electricity price when EV is charging disorderly; θ1 is the basic electricity price adjustment coefficient; P pv,t is the photovoltaic power generation power in period t; is the basic load of the working area during period t.

[0014] In P pv,t = 0, a higher electricity price should be set to reduce the number of charging users during this period; during the photovoltaic power generation period, P pv,t >0, judge the photovoltaic output absorption situation at this time. If the photovoltaic output is fully absorbed, At this time, the electricity price is the initial electricity price c for each period 0,t , if not completely absorbed: The electricity price should be further reduced based on the initial electricity price according to the proportion of photovoltaic power generation exceeding the basic load, so as to attract most users to charge during this period.

[0015] Based on the above ideas, the dynamic electricity price setting mechanism of the work area is as follows:

[0016]

[0017] In formula (2): is the electricity price of the working area during period t; C pv0 is the electricity price when there is no photovoltaic output; C pv1 is the electricity price during the photovoltaic output period; θ pv1 No photovoltaic output price adjustment coefficient; Wload,sum is the total daily base load; W pv,sum is the total daily photovoltaic power generation; θ pv2 It is the photovoltaic power price adjustment coefficient.

[0018] In step 1, the time-of-use pricing strategy divides the 24 hours of the day into peak, flat, and off-peak periods based on the changes in the basic load of the power grid.

[0019]

[0020] In formula (3): is the charging electricity price in the residential area during period t; δ is the fluctuation range of time-of-use electricity price; [t f1 , t f2 ] is the peak period, [t g1 , t g2 ] is the valley period, and other periods are normal periods.

[0021] In the step 2, a demand-price response model for electric vehicles under regional electricity prices is established, including the following steps: Step 2.1: solving the change in charging demand, the change in charging price, and the regional price elasticity coefficient;

[0022] Step 2.2: Based on the above variables, derive the change in the response of the electric vehicle in area j during time period l;

[0023] In step 2.1, after the implementation of regional electricity prices, the response behavior of electric vehicles to electricity prices is not only related to the electricity prices in their region, but also to the electricity prices in other regions. Therefore, the regional price elasticity coefficient is established as shown in formula (4):

[0024]

[0025] In formula (4), l and m represent time period l and time period m respectively, j and h represent area j and area h respectively. When j = h, and are the self-elastic coefficient and mutual elastic coefficient of the same region; when j≠h, l=m, is the elastic coefficient of different regions at the same time; when j≠h, l≠m, is the elasticity coefficient in different regions and time periods. They are the EV charging response when the electric vehicle in region j and time period l is affected by the electricity price in region h and time period m, is the change in electricity price in region h during period m, is the original number of EVs charged in region j during time period l.

[0026] In step 2.2, the electric vehicle user demand-price response model under regional electricity prices can be derived from formula (4):

[0027]

[0028] In formula (5): is the number of electric vehicles charged in region j during time period l after regional electricity price guidance; M is the number of regions; and T is the scheduling period.

[0029] In step 3, for the distribution network, objective function 1 and objective function 2 are constructed with the maximum photovoltaic utilization rate in the working area and the minimum peak-to-valley difference of the equivalent load in the residential area; for electric vehicle users, objective function 3 is constructed with the minimum user charging cost;

[0030] 1) The objective function 1 with the maximum photovoltaic utilization rate in the working area is as follows:

[0031]

[0032] In formula (6): W is the maximum photovoltaic utilization rate in the working area; is the net charging and discharging power of all electric vehicles in the working area during period t; T is the scheduling period; and t is the tth period in the scheduling period.

[0033] in:

[0034]

[0035] In formula (7): is the sum of the number of EVs participating in the scheduling in the work area during period t; P ch,i (t) is the charging power of the i-th EV in period t, P dch,i (t) is the discharge power of the i-th EV during period t; η c is the EV charging coefficient, η d is the EV discharge coefficient.

[0036] The formula for equivalent load in residential areas is:

[0037]

[0038] In formula (8): is the equivalent load of EV and base load in the residential area during period t; is the net charging and discharging power of all EVs in the residential area during period t; is the basic load of the residential area during period t;

[0039]

[0040] In formula (9): is the net charging and discharging power of all EVs in the residential area during period t; is the sum of the number of EVs participating in the dispatch in the residential area during period t.

[0041] 2) The objective function 2, which minimizes the peak-to-valley difference of equivalent load in the residential area, is as follows:

[0042]

[0043] In formula (10): H is the minimum peak-to-valley difference of equivalent load in residential area; is the maximum value of the equivalent load; is the minimum value of equivalent load.

[0044] 3) The objective function 3 with the lowest user charging cost is as follows:

[0045]

[0046] In formula (11), F is the lowest charging and discharging cost for electric vehicle users; Δt is the length of a scheduling period.

[0047] The constraints of the above multi-objective function include photovoltaic output constraints, electric vehicle charging and discharging constraints, and user travel demand state of charge constraints, among which:

[0048] a: Photovoltaic output constraints are:

[0049] 0≤P pv,t ≤P pv,max (12);

[0050] In formula (12): P pv,max It is the maximum output of distributed photovoltaic.

[0051] b: Electric vehicle charging and discharging constraints are:

[0052]

[0053] In formula (13): P i,t is the charging and discharging power of the i-th electric vehicle in the cooperative mode during period t; P ch Charging power for EV; P dis is the EV discharge power; t is the tth period in the scheduling cycle; t i,ar is the time it takes for the i-th vehicle to arrive at the residential area / work area; t i,dep is the time it takes for the i-th vehicle to leave the residential area / work area; S i,t is the state of charge of the i-th electric vehicle at time t; S i,t-1 is the state of charge of the i-th electric vehicle at time t-1; E0 is the rated capacity of the battery; k is the charge and discharge coefficient S i,min is the minimum power of the i-th electric vehicle; S i,maxis the maximum power of the i-th electric vehicle.

[0054] c: The charging state constraint of the user's travel demand is:

[0055] S i,exp ≤S i,dep ≤S i,max (14);

[0056] In formula (14): S i,exp is the expected SOC when the i-th EV finishes charging; S i,dep is the SOC when the i-th EV finishes charging.

[0057] The said step 4 includes the following steps:

[0058] step1: Set initial data for the operating parameters, randomly generate N individuals (initial solution set) in the solution set space, that is, the initial population P t , and use it as the parental population; set the iteration number t = 0; calculate the objective function values f1(x), f2(x), … f m (x) for each individual, where m is the number of objectives;

[0059] step2: Perform non-dominated sorting on the parental population P t , and perform selection, crossover and mutation operations. Generate a first-generation subgroup Q t ;

[0060] Non-dominated sorting: Stratify the individuals in the population according to the dominance relationship. The first layer is all individuals that are not dominated by other individuals. The second layer is individuals that are dominated by the individuals in the first layer but not dominated by other individuals in the same layer. Subsequent layers are类推. The dominance relationship is defined as: For individuals i and j, when and only when for all objectives k, f k (i) ≤ f k (j), and there is at least one objective such that f k (i) < f k (j), as shown in formula (15).

[0061]

[0062] In formula (15): i < j means individual i dominates individual j; represents equivalent to in mathematical logic symbols; f k (i) is the value of individual i on the k-th objective function; f k (j) is the value of individual j on the k-th objective function; represents exists in mathematical logic symbols.

[0063] Selection: When the parent individuals do not belong to the same non-dominated layer, retain the individuals with smaller layer numbers; when the parent individuals belong to the same non-dominated layer, the crowding degree d i Individuals with larger distances are preferred to be retained. k , the crowding degree of individual i is the sum of the absolute values ​​of the function value differences of adjacent individuals under the target, which is normalized as shown in formula (16).

[0064]

[0065] In formula (16): m is the target number; and is the maximum and minimum value of the current layer under the target; f k (i+1) is the value of individual i+1 on the kth objective function; f k (i-1) is the value of individual i-1 on the kth objective function.

[0066] Crossover: Pair the parent individuals in the mating pool and use simulated binary crossover. Input parent individuals x1, x2, crossover probability p c , distribution index η n (usually η n =15), output offspring individuals y1, y2. For each decision variable j, generate a random number u j ∈[0,1], if η n ≤p c ,but:

[0067]

[0068] In formula (17): j To control the degree of deviation between the value after crossover and the parent generation; η n is the cross-distribution index, η n The larger it is, the closer the offspring is to the parent;

[0069] The descendant variable values ​​are:

[0070] y 1j =0.5[(1+β j )x 1j +(1-β j )x 2j ] (18);

[0071] y 2j =0.5[(1-β j )x 1j +(1+β j )x 2j ] (19);

[0072] In the above formula: y 1j is the variable value of the offspring individual y1; y 2j is the variable value of the offspring individual y2; x 1j is the variable value of the parent individual x1; x 2j is the variable value of the parent individual x2.

[0073] Mutation: For each gene position of the offspring, with probability p m Perform polynomial mutation. Input offspring individual y, mutation probability p m , distribution index η m (usually η m =20), variable lower bound y min and upper bound y max For each decision variable j, generate a random number r j ∈[0,1], if r j ≤p m ,but:

[0074]

[0075] In formula (20): j is the degree of deviation between the mutated value and the original value; η m is the variation distribution index.

[0076] The variable value after mutation is:

[0077] y′ j =y j +γ j (y max,j -y min,j ) (twenty one);

[0078] In formula (21): y′ j is the variable value of the offspring individual y after mutation; j is the variable value of the offspring individual y; max,j is the upper bound of the variable value of the offspring individual y; min,j is the lower bound of the variable value of the offspring individual y.

[0079] Step 3: Merge the parent and child generations: t and Q t Fusion to form a temporary population R t , perform fast non-dominated sorting and crowding calculation on the temporary population. The specific formula of fast non-dominated sorting is shown in formula (15). The steps of fast non-dominated sorting are shown in the non-dominated sorting step in step 2, and the crowding calculation is shown in formula (16). The best individual is selected according to the individual's non-dominated sorting layer number and crowding distance. When the temporary population R tWhen the individuals in the population do not belong to the same non-dominated layer, the individuals with smaller stratification numbers are retained; when the individuals belong to the same non-dominated layer, the individuals with larger crowding distance are retained first; and the best individuals selected are used as the parent population P in the next generation evolution operation. t , the number of iterations increases by one.

[0080] Step 4: Determine whether the number of iterations reaches the preset upper limit. If so, end the run; otherwise, jump to step 2.

[0081] The flowchart of the non-dominated sorting genetic algorithm with elite strategy is as follows Figure 2 shown.

[0082] The present invention provides a residential-workspace interactive electric vehicle dynamic pricing method that promotes peak shaving and valley filling and distributed photovoltaic consumption. The technical effects are as follows:

[0083] 1) In step 1 of the present invention, the working area is divided into time periods according to the photovoltaic output and consumption situation, and a dynamic time-of-use electricity price based on the photovoltaic output is formulated, which effectively improves the utilization rate of photovoltaic power generation; in the residential area, the peak, flat and valley time periods are divided according to the basic load, which effectively realizes the transfer of electric vehicle load from peak period to valley period, thereby reducing the peak-valley difference of load.

[0084] 2) Step 2 of the present invention combines demand elasticity theory to establish an electric vehicle user demand-price response model under regional electricity prices to describe the relationship between electricity price changes and EV user electricity consumption. Through the demand-price response model, the electric vehicle load in each region after the regional electricity price guidance can be preliminarily derived.

[0085] 3) Step 3 of the present invention solves the objective function by minimizing the peak-to-valley difference of equivalent load in residential areas and maximizing the photovoltaic utilization rate in working areas on the distribution network side, and minimizing the charging and discharging cost of electric vehicle users on the owner side. This ensures the interests of electric vehicle users while reducing the peak-to-valley difference of equivalent load and increasing the photovoltaic absorption rate.

[0086] 4) In response to the mismatch between electric vehicle load and distributed photovoltaic output in time and space, the electric vehicle dynamic pricing method of the present invention sets different electricity prices in residential areas and work areas, which takes advantage of the difference in regional electricity prices to attract more car owners to stop and charge in the work area. The dynamic electricity price based on photovoltaic output set in the work area guides the EV charging load in the work area to be concentrated in the period when photovoltaic power is not fully absorbed. The time-of-use electricity price set in the residential area attracts EV users to charge in the valley period, achieving a balance of electric vehicle load in the time dimension and the space dimension, effectively promoting the utilization of photovoltaic power generation in the work area and the peak shaving and valley filling of the residential area load. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] The present invention will be further described below with reference to the accompanying drawings and examples:

[0088] Figure 1 Flowchart of an embodiment of the present invention.

[0089] Figure 2 This is the flowchart of the non-dominated sorting genetic algorithm with elitist strategy.

[0090] Figure 3 2 is a simulation system diagram of an embodiment of the present invention.

[0091] Figure 4 This is the base load curve for each area.

[0092] Figure 5 This is the photovoltaic output prediction curve under clear weather.

[0093] Figure 6 This is a schematic diagram of the dynamic electricity price in the working area when the photovoltaic output is sufficient.

[0094] Figure 7 A comparison chart of electric vehicle charging costs under different strategies.

[0095] Figure 8 This is a comparison chart of residential area model optimization results under different electricity prices.

[0096] Figure 9 The figure shows the comparison of optimization results of the work area model under different electricity prices. DETAILED DESCRIPTION

[0097] A method for interactive dynamic pricing of electric vehicles in residential and work areas to promote peak-valley shifting and distributed photovoltaic energy consumption is proposed. The method includes the following steps: For electric vehicle users charging in work areas, a dynamic time-of-use electricity price is established based on the day-ahead predicted photovoltaic output information. For electric vehicle users charging in residential areas, a fixed time-of-use electricity price is established based on the day-ahead predicted residential area base electricity load. A demand-price response model for electric vehicle users under regional electricity prices is established to measure the sensitivity of electric vehicle users' electricity consumption to price fluctuations. A multi-objective function is then established to optimize the scheduling of electric vehicles, with the goals of maximizing photovoltaic utilization in work areas, minimizing the peak-valley difference in equivalent load in residential areas, and minimizing charging costs for electric vehicle users. Finally, a non-dominated sorting genetic algorithm (NSGA-II) with an elitist strategy is used to solve this multi-objective optimization problem. By guiding regional dynamic electricity prices, the present invention addresses the spatiotemporal mismatch between photovoltaic output and electric vehicle load, effectively promoting peak-valley shifting and distributed photovoltaic energy consumption.

[0098] Example:

[0099] The simulation divides the distribution network into three areas according to urban functions: residential, work, and commercial. The commercial area does not connect to electric vehicle loads, and all distributed photovoltaic access locations are located in the work area. The photovoltaic penetration rate is set at a high access rate. Users choose to charge in the work area or residential area based on their travel patterns. Electric vehicle users in the work area choose to charge in the work area during the day, while electric vehicle users in the residential area choose to charge at home at night because they leave early and return late. The simulation verification is performed using an improved IEEE 33-node system. Figure 3 The simulation time period is 00:00-24:00, and each scheduling unit period Δt=15min. The basic load curve of each area is as follows Figure 4 As shown, the photovoltaic output prediction curve is as follows Figure 5 The fixed time-of-use electricity price for electric vehicle charging is shown in Table 1. Among the regional elasticity coefficients, the self-elasticity coefficient is -0.25, the peak-flat and flat-valley cross elasticity coefficients are 0.02, the peak-valley cross elasticity coefficient is 0.04, the elasticity coefficients of different regions in the same period are -0.38, the elasticity coefficients of different regions in the peak-valley period are 0.08, and the elasticity coefficients of different regions in the peak-flat period are 0.02.

[0100] Table 1 Time-of-use electricity prices for electric vehicle charging

[0101]

[0102] Assume that there are 1000 electric vehicles connected in the area, and the charging data are shown in Table 2. The Monte Carlo sampling method is used to obtain the data of EV random access. The EV types are all private cars of the same model. The expected SOC of EV when it is off-grid is 0.85, the battery capacity E0 is 51kWh, the charge and discharge efficiency η is 0.92, the rated charging power and discharge power are both 7kW, the electricity price adjustment coefficient θ1 is 0.3, and there is no photovoltaic output electricity price adjustment coefficient θ pv1 Take 2 / 3, the photovoltaic power output price adjustment coefficient θ pv2 Take 1 / 3.

[0103] Table 2 Random access and off-grid status of electric vehicles

[0104]

[0105] The present invention selects the situation where the photovoltaic output is sufficient under clear weather conditions to simulate and verify the strategy, and sets the simulation results of EV distribution networks in different places under the guidance of disordered charging, time-of-use electricity price, and regional electricity price (the electricity price of the present invention) for comparison. Disordered charging refers to the use of disordered charging in the entire area of ​​the distribution network; time-of-use electricity price refers to the use of time-of-use electricity price in the entire area of ​​the distribution network; regional electricity price refers to the use of fixed time-of-use electricity price in residential areas and the use of dynamic real-time electricity price following photovoltaic output in working areas. When photovoltaic output is sufficient, the electricity price in the working area is as follows: Figure 6 As shown. Figure 6 It can be seen that the electricity price in the working area proposed in this paper can well follow the changes in photovoltaic output, and by setting lower electricity prices during the period when photovoltaic power is not fully absorbed, more electric vehicle users can be attracted to charge during this period.

[0106] The charging cost of EV users guided by different electricity prices is as follows: Figure 7 As shown in the figure, the charging cost for users under time-of-use electricity prices is slightly lower than that under disorderly charging. However, after the regional electricity prices and the strategy of the present invention are implemented, the charging cost for EV users decreases more significantly. This decrease is mainly due to the charging cost of users in residential areas. This is because, after the regional electricity prices are implemented, more EV users choose to charge during the photovoltaic output period in the working area. EVs in the working area can consume photovoltaic output while enjoying lower electricity prices.

[0107] Figure 8 The results of the residential area model optimization under different electricity prices are shown. It can be seen that under disordered charging, vehicle owners immediately charge upon returning to their residential areas, which, combined with the regular load, increases the peak load. With the introduction of time-of-use pricing, the peak load decreases significantly. However, these pricing forces a large number of EV users to charge at the beginning of off-peak hours, creating new peaks during off-peak hours. Compared to time-of-use pricing, regional pricing further reduces the peak load by attracting some EVs to relocate their charging locations.

[0108] To further demonstrate the optimization effect of the proposed strategy in residential areas, Table 3 compares the optimization results for residential areas under different electricity prices. As shown in Table 3, after electric vehicles were guided by time-of-use electricity prices, the peak-to-valley load difference of the residential distribution network decreased from 6405.8kW (when electric vehicles were charging in an unordered manner) to 3732.5kW, and the load standard deviation decreased from 1913.74 to 1137.75. This effectively reduced the peak-to-valley load difference and flattened the load curve. However, with the further increase in the number of electric vehicles in the future, the original load valley period may become the new load peak period. Compared with time-of-use electricity prices, the guidance of regional electricity prices for electric vehicles in residential areas reduced the number of electric vehicles charging in the residential area, and the peak load period decreased from 32697.19kW·h to 30748.56kW·h.

[0109] Table 3 Comparison of optimization results of residential areas under different electricity prices

[0110]

[0111] The optimization results of the workspace model are as follows Figure 9Table 4 compares the optimization results for the work zone under different electricity prices. When EVs are charging unordered, the load is concentrated between 7:00 and 10:00 AM, while the period when PV power cannot be fully absorbed is mainly between 9:00 and 3:00 PM. Therefore, the promotion of PV power absorption under unordered EV charging is limited. After EV users in the work zone were guided by time-of-use electricity prices, the optimization results under time-of-use electricity prices did not change significantly compared to when charging was unordered. Compared to time-of-use electricity price guidance, after considering regional electricity prices, the PV absorption rate increased from 67.98% during unordered EV charging to 79.98%, and the amount of PV absorption blocked decreased from 16,118.75 kW to 10,077.19 kW, significantly increasing PV absorption capacity.

[0112] Table 4 Comparison of work area optimization results under different electricity prices

[0113]

Claims

1. A residential-workspace interactive electric vehicle dynamic pricing method to promote peak shaving and valley filling and distributed photovoltaic consumption, characterized by The following steps are involved: Step 1: For electric vehicle users charging at work areas, a dynamic time-of-use electricity price is set based on the day-ahead forecast of PV output. For electric vehicle users charging in residential areas, a fixed time-of-use electricity price is set based on the day-ahead forecast of the residential area's basic electricity load. Step 2: Establish an electric vehicle user demand-price response model under regional electricity prices to measure the sensitivity of electric vehicle users' electricity consumption to electricity price changes; Step 3: With the goals of maximizing the photovoltaic utilization rate in the work area, minimizing the peak-to-valley difference of the equivalent load in the residential area, and minimizing the charging cost for electric vehicle users, a multi-objective function is established to optimize the scheduling of electric vehicles; Step 4: Use the non-dominated sorting genetic algorithm with elitist strategy to solve the multi-objective optimization problem.

2. The residential-workspace interactive electric vehicle dynamic pricing method for promoting peak load shifting and distributed photovoltaic consumption according to claim 1, characterized in that: In step 1, in order to characterize the relationship between photovoltaic output and electricity price, the basic charging electricity price is first associated with photovoltaic output, and the initial electricity price c for each period of the working area is defined. 0,t for: In formula (1), c0 is the basic charging electricity price, that is, the electricity price when EV is charging disorderly; θ1 is the basic electricity price adjustment coefficient; P pv,t is the photovoltaic power generation power in period t; is the basic load of the working area during period t; In P pv,t = 0, set a higher electricity price to reduce the number of charging users during this period; during the photovoltaic power generation period, P pv,t >0, judge the photovoltaic output absorption situation at this time. If the photovoltaic output is fully absorbed, At this time, the electricity price is the initial electricity price c for each period 0,t , if not completely absorbed: On the basis of the initial electricity price, the electricity price will be further reduced according to the proportion of photovoltaic power generation exceeding the basic load, attracting most users to charge during this period.

3. The residential-workspace interactive electric vehicle dynamic pricing method for promoting peak load shifting and distributed photovoltaic consumption according to claim 2, characterized in that: The dynamic electricity price setting mechanism for the work area is as follows: In formula (2): is the electricity price of the working area during period t; C pv0 is the electricity price when there is no photovoltaic output; C pv1 is the electricity price during the photovoltaic output period; θ pv1 No photovoltaic output price adjustment coefficient; W load,sum is the total daily base load; W pv,sum is the total daily photovoltaic power generation; θ pv2 It is the photovoltaic power price adjustment coefficient.

4. The residential-workspace interactive electric vehicle dynamic pricing method for promoting peak load shifting and distributed photovoltaic consumption according to claim 3, characterized in that: The time-of-use pricing strategy is to divide the 24 hours of the day into peak, flat, and off-peak periods according to the changes in the basic load of the power grid and set prices accordingly: In formula (3): is the charging electricity price in the residential area during period t; δ is the fluctuation range of time-of-use electricity price; [t f1 , t f2 ] is the peak period, [t g1 , t g2 ] is the valley period, and other periods are normal periods.

5. The residential-workspace interactive electric vehicle dynamic pricing method for promoting peak load shifting and distributed photovoltaic consumption according to claim 1, characterized in that: In step 2, establishing an electric vehicle user demand-price response model under regional electricity prices includes the following steps: Step 2.1: Calculate the change in charging demand, charging price, and regional price elasticity coefficient; Step 2.2: Based on the above variables, derive the change in the response of the electric vehicle in region j during time period l.

6. The residential-workspace interactive electric vehicle dynamic pricing method for promoting peak load shifting and distributed photovoltaic consumption according to claim 5, characterized in that: In step 2.1, after the implementation of regional electricity prices, the response behavior of electric vehicles to electricity prices is not only related to the electricity prices in their region, but also to the electricity prices in other regions. Therefore, the regional price elasticity coefficient is established as shown in formula (4): In formula (4), l and m represent time period l and time period m respectively, j and h represent area j and area h respectively; when j = h, and are the self-elastic coefficient and mutual elastic coefficient of the same region; when j≠h, l=m, is the elastic coefficient of different regions at the same time; when j≠h, l≠m, is the elasticity coefficient in different regions and time periods; They are the EV charging response when the electric vehicle in region j and time period l is affected by the electricity price in region h and time period m, is the change in electricity price in region h during period m, is the original number of EVs charged in region j during time period l.

7. The residential-workspace interactive electric vehicle dynamic pricing method for promoting peak load shifting and distributed photovoltaic consumption according to claim 6, characterized in that: In step 2.2, the electric vehicle user demand-price response model under regional electricity prices can be derived from formula (4): In formula (5): is the number of electric vehicles charged in region j during time period l after regional electricity price guidance; M is the number of regions; and T is the scheduling period.

8. The residential-workspace interactive electric vehicle dynamic pricing method for promoting peak load shifting and distributed photovoltaic consumption according to claim 1, characterized in that: In step 3, for the distribution network, objective function 1 and objective function 2 are constructed with the maximum photovoltaic utilization rate in the working area and the minimum peak-to-valley difference of the equivalent load in the residential area; for electric vehicle users, objective function 3 is constructed with the minimum user charging cost; 1) The objective function 1 with the maximum photovoltaic utilization rate in the working area is as follows: In formula (6): W is the maximum photovoltaic utilization rate in the working area; is the net charging and discharging power of all electric vehicles in the working area during period t; T is the scheduling period; t is the tth period in the scheduling period; in: In formula (7): is the sum of the number of EVs participating in the scheduling in the work area during period t; P ch,i (t) is the charging power of the i-th EV in period t, P dch,i (t) is the discharge power of the i-th EV during period t; η c is the EV charging coefficient, η d is the EV discharge coefficient; The formula for equivalent load in residential areas is: In formula (8): is the equivalent load of EV and base load in the residential area during period t; is the net charging and discharging power of all EVs in the residential area during period t; is the basic load of the residential area during period t; In formula (9): is the net charging and discharging power of all EVs in the residential area during period t; is the sum of the number of EVs participating in the dispatch in the residential area during period t; 2) The objective function 2, which minimizes the peak-to-valley difference of equivalent load in the residential area, is as follows: In formula (10): H is the minimum peak-to-valley difference of equivalent load in residential area; is the maximum value of the equivalent load; is the minimum value of equivalent load; 3) The objective function 3 with the lowest user charging cost is as follows: In formula (11), F is the lowest charging and discharging cost for electric vehicle users; Δt is the length of a scheduling period.

9. The residential-workspace interactive electric vehicle dynamic pricing method for promoting peak shaving and valley filling and distributed photovoltaic consumption according to claim 1, characterized in that: The constraints of the multi-objective function include photovoltaic output constraints, electric vehicle charging and discharging constraints, and user travel demand state of charge constraints, among which: a: Photovoltaic output constraints are: 0≤P pv,t ≤P pv,max (12); In formula (12): P pv,max The maximum output of distributed photovoltaics; b: Electric vehicle charging and discharging constraints are: P i,t ∈{P ch ,0,-P dis }t∈[t i,ar ,t i,dep ] S i,min ≤S i,t ≤S i,max In formula (13): P i,t is the charging and discharging power of the i-th electric vehicle in the cooperative mode during period t; P ch Charging power for EV; P dis is the EV discharge power; t is the tth period in the scheduling cycle; t i,ar is the time it takes for the i-th vehicle to arrive at the residential area / work area; t i,dep is the time it takes for the i-th vehicle to leave the residential area / work area; S i,t is the state of charge of the i-th electric vehicle at time t; S i,t-1 is the state of charge of the i-th electric vehicle at time t-1; E0 is the rated capacity of the battery; k is the charge and discharge coefficient S i,min is the minimum power of the i-th electric vehicle; S i,max is the maximum power of the i-th electric vehicle; c: User travel demand state of charge constraint is: S i,exp ≤S i,dep ≤S i,max (14); In formula (14): S i,exp is the expected SOC of the i-th EV when charging is completed; S i,dep is the SOC of the i-th EV when charging is completed.

10. The residential-workspace interactive electric vehicle dynamic pricing method for promoting peak load shifting and distributed photovoltaic consumption according to claim 9, characterized in that: The step 4 comprises the following steps: Step 1: Set the initial data for the operating parameters and randomly generate N individuals in the solution space, that is, the initial population P t , and use it as the parent population; set the number of iterations t = 0; calculate the objective function value f1(x), f2(x), ...f for each individual m (x), where m is the target number; Step 2: For the parent population P t Perform non-dominated sorting, selection, crossover and mutation operations; generate a subgroup Q t ; Non-dominated sorting: The individuals in the population are stratified according to the dominance relationship. The first layer is all individuals that are not dominated by other individuals. The second layer is individuals that are dominated by individuals in the first layer but not dominated by other individuals in the same layer. The subsequent layers are deduced in the same way. The dominance relationship is defined as: for individual i and individual j, if and only if for all targets k, f k (i)≤f k (j), and there is at least one goal that makes f k (i) <f k (j), as shown in formula (15); In formula (15): i < j means that individual i dominates individual j; Indicates the equivalent of mathematical logic symbols; f k (i) is the value of individual i on the kth objective function; f k (j) is the value of individual j on the kth objective function; Represented as existence in mathematical logic symbols; Selection: When the parent individuals do not belong to the same non-dominated layer, retain the individuals with smaller layer numbers; when the parent individuals belong to the same non-dominated layer, the crowding degree d i Individuals with larger distances are preferred to be retained. k , the crowding degree of individual i is the sum of the absolute values ​​of the function value differences of adjacent individuals under the target, which is normalized as shown in formula (16). In formula (16): m is the target number; f k max and f k min is the maximum and minimum value of the current layer under the target; f k (i+1) is the value of individual i+1 on the kth objective function; f k (i-1) is the value of individual i-1 on the kth objective function; Crossover: Pair the parent individuals in the mating pool and use simulated binary crossover; input parent individuals x1, x2, crossover probability p c , distribution index η n , output offspring individuals y1, y2; for each decision variable j, generate a random number u j ∈[0,1], if η n ≤p c ,but: In formula (17): j To control the degree of deviation between the value after crossover and the parent generation; η n is the cross-distribution index, η n The larger it is, the closer the offspring is to the parent; The descendant variable values ​​are: y 1j =0.5[(1+β j )x 1j +(1-β j )x 2j ] (18); y 2j =0.5[(1-β j )x 1j +(1+β j )x 2j ] (19); In the above formula: y 1j is the variable value of the offspring individual y1; y 2j is the variable value of the offspring individual y2; x 1j is the variable value of the parent individual x1; x 2j is the variable value of the parent individual x2; Mutation: For each gene position of the offspring, with probability p m Perform polynomial mutation; input offspring individual y, mutation probability p m , distribution index η m , variable lower bound y min and upper bound y max ; For each decision variable j, generate a random number r j ∈[0,1], if r j ≤p m ,but: In formula (20): j is the degree of deviation between the mutated value and the original value; η m is the variation distribution index; The variable value after mutation is: and' j =and j +γ j (and max,j -and min,j ) (21); In formula (21): y′ j is the variable value of the offspring individual y after mutation; j is the variable value of the offspring individual y; max,j is the upper bound of the variable value of the offspring individual y; min,j is the lower bound of the variable value of the offspring individual y; Step 3: Merge the parent and child generations: t and Q t Fusion to form a temporary population R t , perform fast non-dominated sorting and crowding calculation on the temporary population. The specific formula of fast non-dominated sorting is shown in formula (15). The steps of fast non-dominated sorting are shown in the non-dominated sorting step in step 2, and the crowding calculation is shown in formula (16). The best individual is selected according to the individual's non-dominated sorting layer number and crowding distance. When the temporary population R t When the individuals in the population do not belong to the same non-dominated layer, the individuals with smaller stratification numbers are retained; when the individuals belong to the same non-dominated layer, the individuals with larger crowding distance are retained first; and the best individuals selected are used as the parent population P in the next generation evolution operation. t , the number of iterations increases by one; Step 4: Determine whether the number of iterations reaches the preset upper limit. If so, end the run; otherwise, jump to step 2.

Images