Charge and Discharge Scheduling Method Considering Both Supply and Demand Sides

Abstract

The present invention discloses a charging and discharging scheduling method that takes both supply and demand sides into account. The method steps are as follows: first, the travel characteristics of electric vehicle users are sampled according to the probability density function, the cluster agent obtains the basic information of the power grid and refreshes the information of vehicles connected to the network, and the feasibility of scheduling is judged according to the user's vehicle demand. Then, a charging and discharging priority list is formulated based on the three aspects of reliability, responsiveness and response willingness. Then, the needs of both the supply and demand sides are comprehensively considered, and the battery charging loss factor is taken into account. A two-stage multi-objective optimization model is established with the goals of minimizing the user cost of electric vehicles and minimizing the peak-to-valley difference, so as to obtain the optimization strategy and feedback it to the charging device for execution and update. The present invention fully respects the user's scheduling ability and willingness, arranges the charging and discharging priorities of the dispatchable vehicles, ensures the stability of vehicle charging and discharging, and takes into account the interests of the power grid side and the user side, and has strong practical significance.

Description

Technical Field

[0001] The present invention relates to the technical field of electric vehicles, and in particular to a charging and discharging scheduling method taking both supply and demand sides into consideration. Background Art

[0002] Currently, electric vehicle scheduling methods can be primarily categorized as day-ahead optimization and real-time optimization. Day-ahead scheduling uses historical statistical data to predict load and schedule charging and discharging. However, its results are significantly affected by factors such as vehicle owner preferences, weather conditions, and holidays, which can easily lead to significant deviations between actual and day-ahead scheduling. Real-time optimization can be achieved by analyzing grid loss sensitivity indicators and incorporating power flow calculations and convex optimization algorithms to achieve a real-time solution for electric vehicle charging and discharging scheduling strategies.

[0003] Currently, most of the research on orderly charging and discharging of electric vehicles has problems such as relatively simple optimization objectives and incomplete considerations. In particular, there is a lack of quantitative analysis of user charging and discharging needs, and most of them do not consider the impact of user charging and discharging needs on control strategy design. Summary of the Invention

[0004] Based on the technical problems existing in the background technology, the present invention proposes a charging and discharging scheduling method that takes into account both the supply and demand sides, fully respects the user's scheduling capabilities and willingness, arranges the charging and discharging priorities of the dispatchable vehicles, ensures the stability of vehicle charging and discharging, and takes into account the interests of the grid side and the user side, which has strong practical significance.

[0005] The present invention proposes a charging and discharging scheduling method that takes both supply and demand into account, and the method steps are as follows:

[0006] S1: Gaussian fitting is performed on the user's travel patterns based on travel data to obtain the probability density function of the time when electric vehicles connect to / leave the grid;

[0007] S2: Obtain basic information about the power grid and refresh the information of vehicles connected to the grid;

[0008] S3: Determine the feasibility of vehicle scheduling;

[0009] S4: Develop a charge and discharge priority list based on reliability, responsiveness, and willingness to respond;

[0010] S5: Establish a two-stage optimization model with the goal of minimizing electric vehicle user fees and minimizing peak-to-valley differences and solve the results.

[0011] Preferably, the probability density functions of the time when the electric vehicle connects to / leaves the power grid in S1 are:

[0012]

[0013]

[0014] Among them, f t s (x), μ s , σ s is the probability density function, expectation and variance of the time when electric vehicles are connected to the grid; f t e (x), μ e , σ e is the probability density function, expectation and variance of the time when electric vehicles leave the grid.

[0015] Preferably, the basic information of the power grid obtained in S2 includes the capacity limit of the transformer, the daily load curve of residents and the time-of-use electricity price; the refreshed vehicle information on the grid includes the expected next time of use, the expected remaining power, the willingness to dispatch charge and discharge, the vehicle's remaining charge state and the battery capacity information.

[0016] Preferably, the method steps for determining the feasibility of vehicle scheduling in S3 are as follows:

[0017] S31: Based on the estimated next car usage time provided by the user Get schedulable time Where m represents the mth electric car; e represents leaving the grid; t i Indicates the current moment;

[0018] S32: Calculate the charging time required for the electric vehicle to reach the expected battery capacity based on the information uploaded by the user: where Q q is the expected remaining power, Q s is the vehicle's remaining state of charge, P c is the charging power of the charging pile, η c Charging efficiency of charging piles;

[0019] S33: Determine the feasibility of scheduling. If the following conditions are met: M is the number of electric vehicles connected to the network during this period. If the vehicle meets the scheduling feasibility requirements, it can be scheduled. Otherwise, it will be directly charged at the constant power of the charging pile.

[0020] Preferably, the calculation method of the charging and discharging priority in S4 is:

[0021]

[0022]

[0023]

[0024]

[0025] in, It is charging priority; is the discharge priority; γ j is the coefficient corresponding to the charging mode selected by the user; ω1 and ω2 are the weight coefficients of the reliability and responsiveness indicators in the charging and discharging priority; R m is reliability; f c,m is the response capability; N is the historical number of times the car connects to the smart charging device response strategy; is the actual time that the electric vehicle is off the grid in each response; is the time that electric vehicles enter the grid in each response; P d is the rated discharge power; is the charge and discharge efficiency; Cs m is the battery capacity; is the capacity of the battery at time t; is the initial capacity of the battery.

[0026] Preferably, the charging modes include an immediate charging mode, an orderly charging mode, and an orderly charge and discharge mode, and the corresponding coefficients are 0.1, 0.5, and 1, respectively.

[0027] Preferably, the two-stage optimization of S5 includes a first stage with minimizing user cost as the optimization goal and a second stage with minimizing the peak-to-valley difference in the region as the optimization goal.

[0028] Preferably, the user cost after the optimization in the first stage is:

[0029]

[0030] in, Optimize time windows for sliding; For electric vehicle cluster; x mj is the charge and discharge power of electric vehicle m at time j, with positive values indicating charging and negative values indicating discharging; is the charge and discharge matrix 1 is in charge and discharge state, 0 is in non-charge and discharge state; c j is the electricity price at that time; Δt is the duration of charge and discharge, β and η are the battery loss coefficients; x m(j-1) is the charging and discharging power of electric vehicle m at time j-1.

[0031] Preferably, the peak-to-valley difference after the second stage optimization is: in, are the maximum and minimum values of the total load of the distribution network in the area respectively.

[0032] Preferably, the constraints of the first stage include final SOC constraints, SOC safety constraints, distribution transformer capacity constraints, charging and discharging power constraints, and dispatchable time constraints; the constraints of the second stage include the constraints of the first stage and the user costs after optimization in the first stage.

[0033] Beneficial technical effects of the present invention:

[0034] The charging and discharging scheduling strategy provided by the present invention takes both supply and demand sides into account. Starting from multiple time scales from day-ahead to day-intraday, it fully respects the scheduling capabilities and willingness of users, formulates a charging and discharging priority list based on three aspects: reliability, responsiveness, and response willingness, and ensures the smoothness of vehicle charging and discharging. It establishes a two-stage multi-objective optimization model with the goals of minimizing electric vehicle user costs and minimizing peak-to-valley differences, effectively reducing user costs and significantly reducing the peak-to-valley difference and fluctuation variance of the power grid load. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 This is a schematic diagram of the scheduling technology architecture proposed by the present invention;

[0036] Figure 2 This is a flow chart of the scheduling method proposed by the present invention;

[0037] Figure 3 This is a schematic diagram of the intraday real-time optimization range proposed by the present invention:

[0038] Figure 4 This is the conventional load curve of the power grid proposed by the present invention;

[0039] Figure 5 These are the load curves under different scheduling strategies proposed in this invention. DETAILED DESCRIPTION

[0040] The present invention proposes a charging and discharging scheduling method that takes both supply and demand into account, and the method steps are as follows:

[0041] S1: Gaussian fitting is performed on the user's travel patterns based on the travel data to obtain the probability density function of the time when the electric vehicle is connected to / off the grid. Assuming that the EV user starts charging on the grid after returning home and leaves the grid the next day, the time when the electric vehicle is connected to / off the grid corresponds to the time when the electric vehicle returns home and the time when it leaves home, respectively. The probability density functions of the time when the electric vehicle returns home and the time when the electric vehicle leaves home satisfy the following normal distribution:

[0042]

[0043]

[0044] Among them, f t s (x), μ s , σs is the probability density function, expectation and variance of the time when electric vehicles are connected to the grid; f t e (x), μ e , σ e is the probability density function, expectation and variance of the time when electric vehicles leave the grid.

[0045] S2: Includes transformer capacity limitations, residents' daily load profiles, and time-of-use electricity prices. It updates vehicle information, including the expected next vehicle use time, expected remaining battery charge, charging and discharging dispatchability, vehicle remaining state of charge, and battery capacity information. Discharging dispatchability includes: ① Immediate Charging Mode: EVs are immediately charged at a constant power level at the charging station until the expected charging demand is met; ② Orderly Charging Mode: Cluster agents integrate the user's EV into a scheduled charging schedule; and ③ Orderly Charging and Discharging Mode: Cluster agents integrate the user's EV into a scheduled charging schedule.

[0046] S3: Determine the feasibility of vehicle scheduling. The steps are as follows:

[0047] S31: Based on the estimated next car usage time provided by the user Get schedulable time Where m represents the mth electric car; e represents leaving the grid; t i Indicates the current moment;

[0048] S32: Calculate the charging time required for the electric vehicle to reach the expected battery capacity based on the information uploaded by the user: where Q q is the expected remaining power, Q s is the vehicle's remaining state of charge, P c is the charging power of the charging pile, η c Charging efficiency of charging piles;

[0049] S33: Determine the feasibility of scheduling. If the following conditions are met: M is the number of electric vehicles connected to the network during this period. If the vehicle meets the scheduling feasibility requirements, it can be scheduled. Otherwise, it will be directly charged at the constant power of the charging pile.

[0050] S4: Develop a charge and discharge priority list based on reliability, responsiveness, and willingness to respond. The specific steps are as follows:

[0051] (1) Reliability Where N is the historical number of times the car connects to the smart charging device response strategy; is the actual time that the electric vehicle is off the grid in each response; is the time that electric vehicles enter the grid in each response; Rm It is a historical cumulative mean value, which will be uploaded to the cluster agent for storage. Assume that the value of the first response of all vehicles is 1.

[0052] (2) Responsiveness P d is the rated discharge power; is the charge and discharge efficiency; Cs m is the battery capacity; is the capacity of the battery at time t; is the initial capacity of the battery.

[0053] (3) The response willingness depends on the charging mode selected by the user. According to the three different modes selected by the user, its values are Here the values are 0.1, 0.5, and 1 respectively.

[0054] The calculation formula for discharge priority determined by the three factors of comprehensive reliability, response capability and response willingness is: The charging priority calculation formula is: Among them, ω1 and ω2 are the weight coefficients of reliability and responsiveness indicators in the charge and discharge priority, and both are set to 0.5.

[0055] The priority order of electric vehicle dispatch depends on the c m The electric vehicle charging and discharging priority list formed by the value; for electric vehicle m, when Y c m The higher the value, the higher the EV m The higher the charging priority, the lower the corresponding discharging priority.

[0056] S5: Establish a two-stage optimization model with the goal of minimizing electric vehicle user fees and peak-to-valley differences and solve the results

[0057] The two-stage optimization includes the first stage with the optimization goal of minimizing user cost and the second stage with the optimization goal of minimizing the peak-to-valley difference in the area.

[0058] The user cost after the first phase of optimization is:

[0059]

[0060] in, Optimize time windows for sliding; For electric vehicle cluster; x mj is the charge and discharge power of electric vehicle m at time j, with positive values indicating charging and negative values indicating discharging; is the charge and discharge matrix 1 is in charge and discharge state, 0 is in non-charge and discharge state; c jis the electricity price at that time; Δt is the duration of charge and discharge, β and η are the battery loss coefficients; x m(j-1) is the charging and discharging power of electric vehicle m at time j-1.

[0061] EV cluster Optimize with sliding time window The method for determining is:

[0062] A day is evenly divided into N intervals, each interval is τ, and the charging and discharging power in each interval is assumed to remain unchanged. In order to more accurately obtain the real-time information of electric vehicles connected to the grid, a sliding optimization time window is introduced. To update the EV charging and discharging power in each window. Assume that the current i-th EV cluster is The sliding optimization time window is The current time is t i , the mth electric vehicle EV in the cluster m The entry and end times are and Then when and EV m Belongs to the current cluster And sliding optimization time window

[0063] like Figure 3 As shown, we divide a day into hours, N is 24, and the interval length τ is 1h. Assuming that four electric vehicles EV1, 2, 3, and 4 are connected to the network, then

[0064]

[0065] In addition, the constraints in the first stage include final SOC constraint, SOC safety constraint, distribution transformer capacity constraint, charging and discharging power constraint, and dispatchable time constraint.

[0066] (1) For the final SOC constraint, after the scheduling strategy ends, the final SOC of the electric vehicle needs to meet the next user's vehicle demand, and there needs to be a minimum capacity limit, namely:

[0067]

[0068] Among them, λ m For the minimum battery capacity ratio, The battery capacity of electric vehicles.

[0069] (2) Regarding SOC safety constraints, since overcharging and discharging will cause irreversible damage to the battery, we take into account the limitations of battery safety and capacity. The battery should be kept within the safety constraint range at any charging and discharging moment, that is:

[0070]

[0071] Among them, S min 、S max They are the upper and lower limits of EV's SOC respectively.

[0072] (3) Regarding the capacity constraint of the distribution transformer, at any time, the base load of the area plus the charging and discharging load of all electric vehicles cannot overload the distribution transformer, that is:

[0073]

[0074] Among them, P T The capacity limit of the transformer.

[0075] (4) Regarding the charging and discharging power constraints, electric vehicles have upper and lower limit constraints on charging and discharging power, namely:

[0076]

[0077] Among them, P max is the upper limit of charging power, For charging-only electric vehicles, It is a rechargeable electric vehicle.

[0078] (5) For the dispatchable time constraint, the dispatchable time of electric vehicles is only during the grid connection stage. They are not dispatched before and after the grid connection time. At this time, the charge and discharge decision variables That is, the electric vehicle grid-connected scheduling time meets:

[0079] The peak-to-valley difference after the second stage optimization is: in, are the maximum and minimum values of the total load of the distribution network in the area respectively.

[0080] The constraints of the second stage include the constraints of the first stage and the user costs after optimization in the first stage.

[0081] At this time, the external solver CPLEX can be called through MATLAB's optimization toolbox YALMIP to solve the model. The algorithm flow chart is as follows Figure 2 shown.

[0082] Example

[0083] For example, in a residential area in a certain area, there are 200 electric vehicles within the dispatch range. Smart charging stations use conventional charging mode, with a constant charge and discharge power of 6 kW h and a charge and discharge efficiency of 90%. To ensure charge and discharge safety, the upper and lower SOC limits are 90% and 10%, respectively. The lithium battery capacity of the EVs is 32 kW h, and the power consumption per 100 kilometers is 19.5 kW h.

[0084] Based on the historical data of the power grid in the residential area, we conduct big data analysis and predict the power grid load information of the residential area in each time period of the day, excluding the charging pile, such as Figure 4 shown.

[0085] The present invention is based on the background of time-of-use electricity prices, and the specific electricity price settings are shown in Table 1.

[0086] After adopting the method proposed in this invention, the load curves under different scheduling strategies are as follows: Figure 5 As shown in the figure, the disordered charging strategy is to not schedule the vehicle, and the vehicle is immediately charged to the user's expected power after being connected to the grid, and then disconnected from the grid. The ordered charging strategy is that EV users adopt the charging scheduling strategy of the present invention but do not participate in discharging.

[0087] like Figure 5 As shown, when EVs are charging in an unordered manner, users, considering their own needs, will choose to charge as soon as possible after grid connection to their desired capacity. However, due to their travel habits, users often charge immediately upon returning home, causing their electricity usage to overlap with peak hours in residential areas, resulting in a "peak upon peak" phenomenon. The load peaks at around 6:00 PM, reaching 2282.77 kW. This overloads transformers, potentially creating a dangerous situation, increases the peak-to-valley load difference, impacts grid stability, and increases electricity costs for EV users. When EVs are charging in an organized manner, under the prevailing time-of-use electricity pricing, EV users tend to charge during the load valley period (00:00–08:00). This significantly reduces the peak-to-valley difference and fluctuation variance of the grid load, increasing the load valley to 1697.44 kW, improving grid load and significantly reducing user costs.

[0088] When electric vehicles implement the proposed strategy, compared to the other two strategies, it reduces user electricity costs while effectively achieving peak load shifting and valley filling. User costs are reduced by 53% compared to disordered charging. Furthermore, this strategy reduces peak load to 2142.69 kW and increases valley load to 1697.44 kW, resulting in a peak-to-valley difference of 445.25 kW. This represents a 33.29% reduction compared to disordered charging, smoothing load fluctuations and significantly improving grid load, thus enabling the smooth implementation of this measure.

[0089] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A charging and discharging scheduling method taking into account both supply and demand, characterized in that: The steps are as follows: S1: Gaussian fitting is performed on the user's travel patterns based on travel data to obtain the probability density function of the time when electric vehicles connect to / leave the grid; S2: Obtain basic information about the power grid and refresh the information of vehicles connected to the grid; S3: Determine the feasibility of vehicle scheduling; S4: Develop a charge and discharge priority list based on reliability, responsiveness, and willingness to respond; S5: Establish a two-stage optimization model with the goal of minimizing electric vehicle user fees and peak-to-valley differences and solve the results; The method steps for determining the feasibility of vehicle scheduling in S3 are as follows: S31: Based on the estimated next car usage time provided by the user Get schedulable time Where m represents the mth electric car; e represents leaving the grid; t i Indicates the current moment; S32: Calculate the charging time required for the electric vehicle to reach the expected battery capacity based on the information uploaded by the user: where Q q is the expected remaining power, Q s is the vehicle's remaining state of charge, P c is the charging power of the charging pile, η c Charging efficiency of charging piles; S33: Determine the feasibility of scheduling. If the following conditions are met: m=1,2,…,M, where M is the number of electric vehicles connected to the network during this period. If the vehicle meets the scheduling feasibility requirements, it can be scheduled. Otherwise, it will be directly charged at the constant power of the charging pile.

2. The charge and discharge scheduling method considering both supply and demand sides according to claim 1, characterized in that: The probability density functions of the time when electric vehicles connect to / leave the grid in S1 are: in, μ s , σ s The probability density function, expectation and variance of the time when electric vehicles are connected to the grid; μ e , σ e is the probability density function, expectation and variance of the time when electric vehicles leave the grid.

3. The charge and discharge scheduling method considering both supply and demand sides according to claim 1, characterized in that: The basic grid information obtained in S2 includes the capacity limit of the transformer, the daily load curve of residents and the time-of-use electricity price; the updated vehicle information includes the expected next use time, the expected remaining power, the charging and discharging dispatchability willingness, the vehicle's remaining charge state and the battery capacity information.

4. The charge and discharge scheduling method considering both supply and demand sides according to claim 1, characterized in that: The calculation method of the charge and discharge priority in S4 is: in, It is charging priority; is the discharge priority; γ j is the coefficient corresponding to the charging mode selected by the user; ω1 and ω2 are the weight coefficients of the reliability and responsiveness indicators in the charging and discharging priority; R m is reliability; f c,m is the response capability; N is the historical number of times the car connects to the smart charging device response strategy; is the actual time that the electric vehicle is off the grid in each response; is the time that electric vehicles enter the grid in each response; P d is the rated discharge power; is the charge and discharge efficiency; Cs m is the battery capacity; is the capacity of the battery at time t; is the initial capacity of the battery.

5. The charge and discharge scheduling method considering both supply and demand sides according to claim 4, characterized in that: The charging modes include an immediate charging mode, an orderly charging mode, and an orderly charge and discharge mode, and the corresponding coefficients are 0.1, 0.5, and 1, respectively.

6. The charge and discharge scheduling method considering both supply and demand sides according to claim 1, characterized in that: The two-stage optimization of S5 includes a first stage with the optimization goal of minimizing user cost and a second stage with the optimization goal of minimizing the peak-to-valley difference in the region.

7. The charge and discharge scheduling method considering both supply and demand sides according to claim 6, characterized in that: The user cost after the optimization in the first stage is: in, Optimize time windows for sliding; For electric vehicle cluster; x mj is the charge and discharge power of electric vehicle m at time j, with positive values indicating charging and negative values indicating discharging; is the charge and discharge matrix 1 is in charge and discharge state, 0 is in non-charge and discharge state; c j is the electricity price at that time; Δt is the duration of charge and discharge, β and η are the battery loss coefficients; x m(j-1) is the charging and discharging power of electric vehicle m at time j-1.

8. The charge and discharge scheduling method considering both supply and demand sides according to claim 6, characterized in that: The peak-to-valley difference after the second stage optimization is: in, are the maximum and minimum values of the total load of the distribution network in the area respectively.

9. The charge and discharge scheduling method considering both supply and demand sides according to claim 1, characterized in that: The constraints in the first stage include final SOC constraints, SOC safety constraints, distribution transformer capacity constraints, charging and discharging power constraints, and dispatchable time constraints; the constraints in the second stage include the constraints in the first stage and the user costs after optimization in the first stage.

Images