A real-time dispatching method for electric vehicles considering forecasted load and user demand

Abstract

This invention discloses a real-time dispatching method for electric vehicles that takes into account predicted load and user demand. The method includes the following steps: categorizing electric vehicle users based on whether they have signed a contract with an electric vehicle aggregator and differences in charging needs, and establishing charging control models for each type of electric vehicle user accordingly; establishing a relationship model between the response probability of non-contracted users and the compensation electricity price; establishing a subsidy mechanism for the power grid to aggregators, and vice versa; calculating the response potential of each type of electric vehicle user, determining the grid-side dispatching demand for each time period based on the predicted load, and formulating a real-time dispatching plan for each time period based on the grid-side dispatching demand and the electric vehicle user potential; and establishing a real-time optimization model and using a particle swarm algorithm to solve for electric vehicle charging power. This invention can achieve good load peak shaving while meeting the different charging needs of electric vehicle users, and can effectively improve power grid security.

Description

Technical Field

[0001] The present invention belongs to the technical field of power system demand response, and in particular relates to a real-time dispatching method for electric vehicles taking into account predicted load and user demand. Background Art

[0002] As the number of electric vehicles connected to the grid continues to increase, the randomness and volatility of their charging load will place a significant operational burden on the power system. It is necessary to enable electric vehicles to participate in distribution network scheduling through orderly charging to reduce load pressure and eliminate the risk of grid overload. Currently, methods for optimizing electric vehicle charging power are divided into day-ahead global optimization and real-time local optimization. Due to the nature of electric vehicles as mobile loads, their arrival and departure times are random, and the battery state of charge at the time of connection has a large degree of uncertainty. Day-ahead scheduling often fails to meet the charging needs of some electric vehicles, necessitating real-time scheduling of electric vehicle users. At the same time, the participation of electric vehicles in coordinated dispatching is a demand response process. At present, demand response is mainly divided into price-based demand response and incentive-based demand response. Under price-based demand response, users independently choose the charging period, and the charging power is not regulated. It cannot fully utilize the demand response potential of electric vehicles and may cause "peak-valley inversion" phenomenon (Cheng Shan, Chen Ziming, Xu Kangyi, et al. Orderly charging and discharging method of electric vehicles based on cooperative game and dynamic time-of-use electricity price [J]. Power System Protection and Control, 2020, 48(21):21-27.); Incentive-based demand response is based on the signing of contracts or agreements. According to the characteristics of different users, incentive strategies such as economic compensation or electricity price discounts can be used to guide users to participate in load adjustment, which can make more full use of the response potential of electric vehicles. Current real-time optimization research is based on the premise that electric vehicles can participate in scheduling or only uses contracted users as demand response targets. It rarely considers the differences in user charging needs and non-contracted users who are unwilling to sign long-term incentive agreements due to the high flexibility of their daily travel needs (Chen Lüpeng, Pan Zhenning, Yu Tao, Wang Keying. Real-time optimization scheduling of large-scale electric vehicles based on dynamic non-cooperative game [J]. Automation of Electric Power Systems, 2019, 43(24): 32-40+66.). Few technologies consider scheduling of predicted loads in real-time optimization, resulting in insufficient real-time scheduling effects and insufficient utilization of the response potential of electric vehicles (Zhang Bingxu, Xu Gang. Rolling time domain optimization of electric vehicle grid connection considering demand differences [J]. Automation of Electric Power Systems, 2020, 44(13): 106-114.). Summary of the Invention

[0003] To address the shortcomings of the existing technology, the present invention proposes to classify electric vehicle users according to whether they have signed a contract and the differences in their charging needs, and based on this, establish charging models for various types of electric vehicles. It introduces a relationship model between the response probability of non-contracted users and the compensation electricity price, establishes a subsidy mechanism for both parties of demand response, and proposes a real-time optimization method for electric vehicle charging power based on scheduling needs and predicted scheduling potential of various types of users, with the increase in the profit ratio of the aggregator and the increase in the average revenue of users as the comprehensive goals.

[0004] The purpose of the present invention is achieved by at least one of the following technical solutions.

[0005] A real-time dispatching method for electric vehicles taking into account predicted load and user demand comprises the following steps:

[0006] S1. Classify electric vehicle users based on whether they have signed a contract with an electric vehicle aggregator and their charging needs, and establish a charging control model for each type of electric vehicle user accordingly;

[0007] S2. Establish a relationship model between the response probability of non-contracted users and the compensation electricity price;

[0008] S3. Establish a subsidy mechanism for the grid to subsidize aggregators and for aggregators to subsidize electric vehicle users participating in demand response;

[0009] S4. Calculate the response potential of various types of electric vehicle users, determine the grid-side dispatching demand for each time period based on the predicted load, and formulate a real-time dispatching plan for each time period based on the grid-side dispatching demand and the potential of electric vehicle users;

[0010] S5. Establish a real-time optimization model, take the grid-side control target power as the constraint, take the profit increase ratio of electric vehicle aggregators and the average revenue increase ratio of electric vehicle users as the comprehensive objectives, and use the particle swarm algorithm to solve the electric vehicle charging power.

[0011] Furthermore, in step S1, the lithium battery is taken as the object, the battery self-discharge process is ignored, and it is approximately assumed that the battery maintains a constant charging power in each optimization period, and a charging model for the electric vehicle is established, which is as follows:

[0012]

[0013] Among them, S ne is the charging power requirement of the electric vehicle; S0 is the initial state of charge (SOC) when the electric vehicle is connected; S ex is the expected state of charge of the electric vehicle when it leaves; C0 is the battery capacity; S(t) is the state of charge of the electric vehicle at time t; η is the charging efficiency; P c (t) is the charging power of the electric vehicle at time t; Δt is the time interval; P cNis the rated charging power of the electric vehicle; t ar With t ex are the arrival time and estimated departure time of the electric vehicle, respectively;

[0014] Determine the power boundary of the electric vehicle as follows:

[0015] T cmin =C0(S ex -S0) / (ηP cN ) (35)

[0016] t ml =t ar +T cmin (36)

[0017]

[0018] t ex =t ar +T tl (38)

[0019] t mc =t ex -T cmin (39)

[0020]

[0021] Among them, T cmin The shortest charging time for electric vehicles; t ml is the fastest departure time of the electric vehicle; S max (t) is the upper limit of the electric vehicle's power at time t in the station; T tl is the length of time the electric vehicle stays in the station; t mc The latest time to start charging the electric vehicle; S min (t) is the lower bound of the electric vehicle’s charge at time t when it is in the station.

[0022] Furthermore, in step S1, it is assumed that the contracted user will actively participate in demand response when there is no urgent travel demand, and set a scheduling power difference limit S dm As a restriction condition for the aggregator to freely dispatch contracted users, it is used to prevent contracted users without urgent travel needs from being in a dispatched state and unable to fully charge. That is, when the difference between the current power of a contracted user and the power charged to the current moment according to the rated power after entering the station exceeds the dispatch power difference limit S dm , then the electric vehicle will no longer participate in the dispatch; the contracted users without urgent travel needs are recorded as Class A users. The charging control model for Class A users is as follows:

[0023]

[0024] Where q(i) is the category of electric vehicle i; d(i) is the dispatch priority of electric vehicle i, d(i) = 1 indicates the highest dispatch priority, d(i) = 2 and d(i) = 3 indicate the dispatch priority is lower in turn; S ex (i) is the expected power of electric vehicle i; S M The maximum charging capacity of the electric vehicle; P min (i,t) and P max (i, t) are the lower and upper limits of the charging power of electric vehicle i at time t; s(i, t) is the dispatchable state of electric vehicle i at time t, s(i, t) = 0 means it is not dispatchable, s(i, t) = 1 means it is dispatchable. Contracted users with urgent travel needs are always in the non-dispatchable state, as follows:

[0025]

[0026] S b (i,t)=S0(i)+ηP cN (tt ar ) / C0

[0027] S d (i,t)=S(i,t)-S b (i,t) (43)

[0028] Among them, S b (i, t) is the charging power of electric vehicle i at rated charging power P cN Time from arrival to destination t ar The standard power charged to time t; S d (i, t) is the dispatch power difference of electric vehicle i at time t; S(i, t) is the power of electric vehicle i at time t; S0(i) is the initial state of charge when electric vehicle i is connected;

[0029] For users who have not signed contracts with aggregators, that is, ordinary users, they often do not choose long-term contracts due to the high flexibility of their daily travel needs. If ordinary users have low travel needs when the aggregator issues a demand response, and hope to participate in demand response in exchange for certain benefits, they can participate in demand response by signing a temporary contract and accept the aggregator's dispatch. Therefore, for ordinary users who are willing to participate in demand response, if their stay time at the charging station exceeds the minimum charging time, they can participate in demand response without reducing their desired charging power. Such users are recorded as Class B users, and their scheduling priority is second only to Class A users. When Class A users cannot meet the grid dispatch requirements, the aggregator needs to use the compensation price for Class B users to guide Class B users to participate in demand response. For ordinary users whose stay time at the charging station is less than the minimum charging time, they need to participate in demand response with the minimum charging power as their charging demand. They are recorded as Class C users, and their scheduling priority is lower than that of Class A and Class B users. When Class A and Class B users cannot meet the grid dispatch requirements, the aggregator needs to pay a higher compensation price than the compensation price for Class B users to guide Class C users to participate in demand response. The charging control model for ordinary users is as follows:

[0030]

[0031] Among them, S zd (i) The minimum charging power provided to electric vehicle i, which is usually lower than the maximum charging power S of the electric vehicle M ;S ex (i) is the expected state of charge of electric vehicle i when it leaves; T cmin (i) is the shortest charging time of electric vehicle i; T tl (i) is the length of time that electric vehicle i stays at the station; S min (i, t) is the lower bound of the amount of electricity consumed by electric vehicle i at the station at time t, S max (i,t) is the upper bound of the amount of electricity that electric vehicle i can hold at the station at time t.

[0032] Furthermore, in step S2, a relationship model between the response probability of non-contracted users and the compensation electricity price is established, specifically as follows:

[0033] When the dispatching potential of contracted users cannot meet the grid demand, it is necessary to formulate a reasonable compensation price to guide ordinary users to participate in the response. Since the higher the compensation price, the higher the benefits of users participating in demand response and the greater the response probability, a direct proportional function is used to establish the relationship between the response probability of non-contracted users, i.e. ordinary users, and the compensation price, as follows:

[0034]

[0035] μ=1 / (c m -c0) (46)

[0036] Among them, p x (t) is the response probability of ordinary users at time t; c(t) is the compensation price at time t; μ is the response probability of ordinary users increased by increasing the unit compensation price; c0 and c m They are the minimum and maximum compensation electricity prices that aggregators provide to ordinary users respectively.

[0037] Furthermore, step S3 includes the following steps:

[0038] S3.1. Establish a compensation mechanism for the power grid to aggregators. When aggregators participate in demand response, the power grid needs to compensate the aggregators based on the degree of load reduction, as follows:

[0039] B g (t) = W y (t) b EVA ·a (47)

[0040]

[0041] D f (t)=W(t) / P ne (t) (49)

[0042]

[0043] Among them, B g (t) is the compensation paid by the power grid to the aggregator at time t; W y (t) is the effective response power of the aggregator at time t; b EVA is the subsidy standard for participating in the response, usually ranging from 0 to 5 (yuan / kW·h); a is the response coefficient, which is 3 under the real-time peak-shaving demand response; D f (t) is the response completion rate of the aggregator at time t; N s (t) is the sum of the number of electric vehicles that have arrived and the number of electric vehicles expected to arrive participating in demand response at time t; P(i,t) is the charging power of electric vehicle i at time t;

[0044] S3.2. Establish a compensation mechanism for electric vehicle users by aggregators, subsidizing electric vehicle users based on the reduced power consumption of electric vehicles, as follows:

[0045]

[0046] B ev (i,t)=c(t)·W(i,t) (52)

[0047] W(i,t)=Δt·(P cN -P(i,t)) (53)

[0048] Among them, B EV (t) is the total subsidy cost of the aggregator to all electric vehicles at time t; B ev (i,t) is the compensation fee paid by the aggregator to electric vehicle i at time t; W(i,t) is the response power of electric vehicle i at time t.

[0049] Furthermore, step S4 includes the following steps:

[0050] S4.1. The dispatch potential of each type of electric vehicle and the dispatch demand in each time period are calculated as follows:

[0051] P ne (t) = P pev (t)+P pb (t)-P aim (54)

[0052] P pev (t) = P cN (N n (t)+N p (t)) (55)

[0053]

[0054] Among them, P ne (t) is the scheduling demand at time t, P pev (t) is the electric vehicle load predicted to be connected before the next moment at moment t, P pb (t) is the normal load predicted at time t, P aim is the control target power on the grid side; N n (t) is the number of electric vehicles at the charging station at time t; N p (t) is the number of electric vehicles expected to arrive at the charging station before the next moment; P cap (q,t) is the dispatch potential of q-type electric vehicles at time t; N s (q,t) is the number of electric vehicles of type q at time t, q = A, B or C;

[0055] S4.2. Develop real-time dispatch plans for each period based on the dispatch demand of the grid side and user potential. When there is a dispatch demand P in a certain period, ne (t), various types of electric vehicles are dispatched in descending order of dispatch priority. When the dispatch potential of high-priority users cannot meet the dispatch demand of the power grid, the electric vehicles of the lower level are dispatched.

[0056] Furthermore, in step S4.2, the specific scheduling scheme is as follows:

[0057] 1) When P ne (t) <Pcap (A, t), some Class A electric vehicles are dispatched within time t;

[0058] 2) When P cap (A,t)≤P ne (t) <P cap (A,t)+P cap (B, t), all Class A electric vehicles and some Class B electric vehicles are dispatched within time t;

[0059] 3) When P cap (A,t)+P cap (B,t)≤P ne (t) <P cap (A,t)+P cap (B,t)+P cap (C,t), dispatches all A and B class electric vehicles and some C class electric vehicles at time t;

[0060] 4) When P cap (A,t)+P cap (B,t)+P cap (C,t)≤P ne At time (t), all electric vehicles are dispatched within time t.

[0061] Furthermore, S5 includes the following steps:

[0062] S5.1. The real-time optimization model uses the grid-side control target power as a constraint:

[0063]

[0064] Among them, P(i,t) is the charging power of electric vehicle i after scheduling at time t; P u (t) is the sum of the non-dispatchable loads at time t, including the electric vehicle load and conventional load in the non-dispatchable state; P aim is the control target power on the grid side;

[0065] S5.2. The real-time optimization model takes the profit growth ratio of electric vehicle aggregators and the average revenue growth ratio of electric vehicle users as comprehensive objectives, as follows:

[0066] The aggregator’s costs include the loss of service fee revenue ΔB at time t ser (t) and compensation to users B EV (t), the benefit is the compensation fee B from the power grid company at time t g (t), that is:

[0067]

[0068] B EVA (t) = B g (t)-ΔB ser (t)-B EV (t) (59)

[0069] η EVA (t) = B EVA (t) / B ser (t) (60)

[0070] Among them, B ser (t) and B' ser (t) is the service fee income expected before and after optimization at time t; P(t) is the total charging load expected after optimization at time t; B EVA (t) is the expected profit of the aggregator after optimization at time t; c ser (t) is the charging service price at time t; η EVA (t) is the expected profit growth ratio of the aggregator after optimization at time t;

[0071] For electric vehicle users, their costs mainly include response costs. Generally, user response costs have a monotonic and concave characteristic with respect to power reduction, so they are represented by a quadratic function:

[0072] B x (i,t)=a x [W(i,t)] 2 +b x W(i,t) (61)

[0073] Among them, B x (i, t) is the response cost of electric vehicle user i at time t; W(i, t) is the reduced power consumption of electric vehicle user i at time t; a x and b x are coefficients, all of which are constants greater than 0;

[0074] User benefits include the reduced electricity cost ΔB of electric vehicle user i at time t c (i, t) and the compensation fee B paid by the aggregator to electric vehicle user i at time t ev (i,t), that is:

[0075]

[0076] B evs (i,t)=B ev (i,t)+ΔB c (i,t)-B x (i,t) (63)

[0077] η evs(i,t)=B evs (i,t) / B c (i,t) (64)

[0078] Among them, B c (i,t) is the electricity cost of electric vehicle user i at time t, c ch (t) is the charging price at time t; B evs (i,t),η evs (i, t) are the benefits and benefit increase ratio of electric vehicle user i participating in demand response at time t;

[0079] Comprehensively examine the average benefit B of all users participating in demand response at time t evm (t) and average return ratio η evm (t) is:

[0080]

[0081] In order to balance the needs and interests of all parties, the grid-side control target power is used as the dispatch constraint, and the maximization of the aggregator's profit increase ratio and the user's average revenue increase ratio is used as the comprehensive optimization goal, namely:

[0082] maxF(t)=β1η EVA (t)+β2η evm (t) (66)

[0083] Among them, β1 and β2 are the weight coefficients of the aggregator's profit growth ratio and the user's average revenue ratio respectively.

[0084] The beneficial effects of the present invention are:

[0085] (1) Users can gain benefits by participating in demand response; the peak load of the distribution network is significantly improved; the power grid company has achieved improvements in grid security at a certain economic cost, and its need to implement demand response has been met.

[0086] (2) The electric vehicle user classification method of the present invention is based on whether the user has signed an incentive agreement with the aggregator and the user's sufficient time to stay at the charging station. It can ensure the user's travel needs and charging needs, fully utilize the response potential of electric vehicle users, and achieve demand response with good peak-shaving effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] Figure 1 This is a flow chart of a real-time dispatching method for electric vehicles taking into account predicted load and user demand in an embodiment of the present invention.

[0088] Figure 2 1 is a graph showing conventional load in a commercial area and predicted load for electric vehicles in an embodiment of the present invention.

[0089] FIG3 is a diagram showing the response effect of real-time scheduling in an embodiment of the present invention, wherein: Figure 3a Schematic diagram of the total load curve before and after response. Figure 3b Schematic diagram of electric vehicle power curve before and after response, Figure 3c Schematic diagram of the number of responding vehicles and average power.

[0090] Figure 4 This is a load response curve diagram showing whether a common user responds or not in an embodiment of the present invention.

[0091] Figure 5 3 is a load curve comparison diagram of different proportions of contracted users in an embodiment of the present invention. DETAILED DESCRIPTION

[0092] The accompanying drawings are for illustrative purposes only and are not to be construed as limiting the present invention. To better illustrate the present embodiment, some components in the accompanying drawings may be omitted, enlarged, or reduced in size, and do not represent actual product dimensions. Those skilled in the art will appreciate that some well-known structures and their descriptions may be omitted from the accompanying drawings. The positional relationships depicted in the accompanying drawings are for illustrative purposes only and are not to be construed as limiting the present invention.

[0093] Example 1:

[0094] A real-time dispatching method for electric vehicles taking into account the forecast load and user demand, such as Figure 1 As shown, the following steps are included:

[0095] S1. Classify electric vehicle users based on whether they have signed a contract with an electric vehicle aggregator and their charging needs, and establish a charging control model for each type of electric vehicle user accordingly;

[0096] Taking lithium batteries as the object, ignoring the battery self-discharge process and approximately assuming that the battery maintains a constant charging power in each optimization period, a charging model for electric vehicles is established, as follows:

[0097]

[0098] Among them, S ne is the charging power requirement of the electric vehicle; S0 is the initial state of charge (SOC) when the electric vehicle is connected; S ex is the expected state of charge of the electric vehicle when it leaves; C0 is the battery capacity; S(t) is the state of charge of the electric vehicle at time t; η is the charging efficiency; P c (t) is the charging power of the electric vehicle at time t; Δt is the time interval; P cN is the rated charging power of the electric vehicle; t ar With t exare the arrival time and estimated departure time of the electric vehicle, respectively;

[0099] Determine the power boundary of the electric vehicle as follows:

[0100] T cmin =C0(S ex -S0) / (ηP cN ) (68)

[0101] t ml =t ar +T cmin (69)

[0102]

[0103] t ex =t ar +T tl (71)

[0104] t mc =t ex -T cmin (72)

[0105]

[0106] Among them, T cmin The shortest charging time for electric vehicles; t ml is the fastest departure time of the electric vehicle; S max (t) is the upper limit of the electric vehicle's power at time t in the station; T tl is the length of time the electric vehicle stays in the station; t mc The latest time to start charging the electric vehicle; S min (t) is the lower bound of the electric vehicle’s charge at time t when it is in the station.

[0107] Assume that the contracted user will actively participate in demand response when there is no urgent travel demand, and set a scheduling power difference limit S dm As a restriction condition for the aggregator to freely dispatch contracted users, it is used to prevent contracted users without urgent travel needs from being in a dispatched state and unable to fully charge. That is, when the difference between the current power of a contracted user and the power charged to the current moment according to the rated power after entering the station exceeds the dispatch power difference limit S dm , then the electric vehicle will no longer participate in the dispatch; the contracted users without urgent travel needs are recorded as Class A users. The charging control model for Class A users is as follows:

[0108]

[0109] Where q(i) is the category of electric vehicle i; d(i) is the dispatch priority of electric vehicle i, d(i) = 1 indicates the highest dispatch priority, d(i) = 2 and d(i) = 3 indicate the dispatch priority is lower in turn; S ex (i) is the expected power of electric vehicle i; S M The maximum charging capacity of the electric vehicle; P min (i,t) and P max (i, t) are the lower and upper limits of the charging power of electric vehicle i at time t; s(i, t) is the dispatchable state of electric vehicle i at time t, s(i, t) = 0 means it is not dispatchable, s(i, t) = 1 means it is dispatchable. Contracted users with urgent travel needs are always in the non-dispatchable state, as follows:

[0110]

[0111]

[0112] Among them, S b (i, t) is the charging power of electric vehicle i at rated charging power P cN Time from arrival to destination t ar The standard power charged to time t; S d (i, t) is the dispatch power difference of electric vehicle i at time t; S(i, t) is the power of electric vehicle i at time t; S0(i) is the initial state of charge when electric vehicle i is connected;

[0113] For users who have not signed contracts with aggregators, that is, ordinary users, they often do not choose long-term contracts due to the high flexibility of their daily travel needs. If ordinary users have low travel needs when the aggregator issues a demand response, and hope to participate in demand response in exchange for certain benefits, they can participate in demand response by signing a temporary contract and accept the aggregator's dispatch. Therefore, for ordinary users who are willing to participate in demand response, if their stay time at the charging station exceeds the minimum charging time, they can participate in demand response without reducing their desired charging power. Such users are recorded as Class B users, and their scheduling priority is second only to Class A users. When Class A users cannot meet the grid dispatch requirements, the aggregator needs to use the compensation price for Class B users to guide Class B users to participate in demand response. For ordinary users whose stay time at the charging station is less than the minimum charging time, they need to participate in demand response with the minimum charging power as their charging demand. They are recorded as Class C users, and their scheduling priority is lower than that of Class A and Class B users. When Class A and Class B users cannot meet the grid dispatch requirements, the aggregator needs to pay a higher compensation price than the compensation price for Class B users to guide Class C users to participate in demand response. The charging control model for ordinary users is as follows:

[0114]

[0115] Among them, S zd (i) The minimum charging power provided to electric vehicle i, which is usually lower than the maximum charging power S of the electric vehicle M ;S ex (i) is the expected state of charge of electric vehicle i when it leaves; T cmin (i) is the shortest charging time of electric vehicle i; T tl (i) is the length of time that electric vehicle i stays at the station; S min (i, t) is the lower bound of the amount of electricity consumed by electric vehicle i at the station at time t, S max (i,t) is the upper bound of the amount of electricity that electric vehicle i can hold at the station at time t.

[0116] S2. Establish a relationship model between the response probability of non-contracted users and the compensation electricity price, as follows:

[0117] When the dispatching potential of contracted users cannot meet the grid demand, it is necessary to formulate a reasonable compensation price to guide ordinary users to participate in the response. Since the higher the compensation price, the higher the benefits of users participating in demand response and the greater the response probability, a direct proportional function is used to establish the relationship between the response probability of non-contracted users, i.e. ordinary users, and the compensation price, as follows:

[0118]

[0119] μ=1 / (c m -c0) (79)

[0120] Among them, p x (t) is the response probability of ordinary users at time t; c(t) is the compensation price at time t; μ is the response probability of ordinary users increased by increasing the unit compensation price; c0 and c m They are the minimum and maximum compensation electricity prices that aggregators provide to ordinary users respectively.

[0121] S3. Establish a subsidy mechanism for the grid to subsidize aggregators and for aggregators to subsidize electric vehicle users participating in demand response, including the following steps:

[0122] S3.1. Establish a compensation mechanism for the power grid to aggregators. When aggregators participate in demand response, the power grid needs to compensate the aggregators based on the degree of load reduction, as follows:

[0123] B g (t) = W y (t) b EVA ·a (80)

[0124]

[0125] D f (t)=W(t) / Pne (t) (82)

[0126]

[0127] Among them, B g (t) is the compensation paid by the power grid to the aggregator at time t; W y (t) is the effective response power of the aggregator at time t; b EVA is the subsidy standard for participating in the response, usually ranging from 0 to 5 (yuan / kW·h); a is the response coefficient, which is 3 under the real-time peak-shaving demand response; D f (t) is the response completion rate of the aggregator at time t; N s (t) is the sum of the number of electric vehicles that have arrived and the number of electric vehicles expected to arrive participating in demand response at time t; P(i,t) is the charging power of electric vehicle i at time t;

[0128] S3.2. Establish a compensation mechanism for electric vehicle users by aggregators, subsidizing electric vehicle users based on the reduced power consumption of electric vehicles, as follows:

[0129]

[0130] B ev (i,t)=c(t)·W(i,t) (85)

[0131] W(i,t)=Δt·(P cN -P(i,t)) (86)

[0132] Among them, B EV (t) is the total subsidy cost of the aggregator to all electric vehicles at time t; B ev (i,t) is the compensation fee paid by the aggregator to electric vehicle i at time t; W(i,t) is the response power of electric vehicle i at time t.

[0133] S4. Calculate the response potential of various types of electric vehicle users, determine the grid-side dispatching demand for each time period based on the predicted load, and formulate a real-time dispatching plan for each time period based on the grid-side dispatching demand and the electric vehicle user potential, including the following steps:

[0134] S4.1. The dispatch potential of each type of electric vehicle and the dispatch demand in each time period are calculated as follows:

[0135] P ne (t) = P pev (t)+P pb (t)-P aim (87)

[0136] P pev (t) = PcN (N n (t)+N p (t)) (88)

[0137]

[0138] Among them, P ne (t) is the scheduling demand at time t, P pev (t) is the electric vehicle load predicted to be connected before the next moment at moment t, P pb (t) is the normal load predicted at time t, P aim is the control target power on the grid side; N n (t) is the number of electric vehicles at the charging station at time t; N p (t) is the number of electric vehicles expected to arrive at the charging station before the next moment; P cap (q,t) is the dispatch potential of q-type electric vehicles at time t; N s (q,t) is the number of electric vehicles of type q at time t, q = A, B or C;

[0139] S4.2. Develop real-time dispatch plans for each period based on the dispatch demand of the grid side and user potential. When there is a dispatch demand P in a certain period, ne (t), various types of electric vehicles are dispatched in descending order of dispatch priority. When the dispatch potential of high-priority users cannot meet the dispatch demand of the power grid, electric vehicles at the lower level are dispatched. The specific dispatch plan is as follows:

[0140] 1) When P ne (t) <P cap (A, t), some Class A electric vehicles are dispatched within time t;

[0141] 2) When P cap (A,t)≤P ne (t) <P cap (A,t)+P cap (B, t), all Class A electric vehicles and some Class B electric vehicles are dispatched within time t;

[0142] 3) When P cap (A,t)+P cap (B,t)≤P ne (t) <P cap (A,t)+P cap (B,t)+P cap (C,t), dispatches all A and B class electric vehicles and some C class electric vehicles at time t;

[0143] 4) When P cap(A,t)+P cap (B,t)+P cap (C,t)≤P ne At time (t), all electric vehicles are dispatched within time t.

[0144] S5. Establish a real-time optimization model, using the grid-side control target power as a constraint, the profit increase ratio of electric vehicle aggregators and the average revenue increase ratio of electric vehicle users as comprehensive objectives, and use the particle swarm algorithm to solve the electric vehicle charging power, including the following steps:

[0145] S5.1. The real-time optimization model uses the grid-side control target power as a constraint:

[0146]

[0147] Among them, P(i,t) is the charging power of electric vehicle i after scheduling at time t; P u (t) is the sum of the non-dispatchable loads at time t, including the electric vehicle load and conventional load in the non-dispatchable state; P aim is the control target power on the grid side;

[0148] S5.2. The real-time optimization model takes the profit growth ratio of electric vehicle aggregators and the average revenue growth ratio of electric vehicle users as comprehensive objectives, as follows:

[0149] The aggregator’s costs include the loss of service fee revenue ΔB at time t ser (t) and compensation to users B EV (t), the benefit is the compensation fee B from the power grid company at time t g (t), that is:

[0150]

[0151] B EVA (t) = B g (t)-ΔB ser (t)-B EV (t) (92)

[0152] η EVA (t) = B EVA (t) / B ser (t) (93)

[0153] Among them, B ser (t) and B' ser (t) is the service fee income expected before and after optimization at time t; P(t) is the total charging load expected after optimization at time t; B EVA (t) is the expected profit of the aggregator after optimization at time t; c ser(t) is the charging service price at time t; η EVA (t) is the expected profit growth ratio of the aggregator after optimization at time t;

[0154] For electric vehicle users, their costs mainly include response costs. Generally, user response costs have a monotonic and concave characteristic with respect to power reduction, so they are represented by a quadratic function:

[0155] B x (i,t)=a x [W(i,t)] 2 +b x W(i,t) (94)

[0156] Among them, B x (i, t) is the response cost of electric vehicle user i at time t; W(i, t) is the reduced power consumption of electric vehicle user i at time t; a x and b x are coefficients, all of which are constants greater than 0;

[0157] User benefits include the reduced electricity cost ΔB of electric vehicle user i at time t c (i, t) and the compensation fee B paid by the aggregator to electric vehicle user i at time t ev (i,t), that is:

[0158]

[0159] B evs (i,t)=B ev (i,t)+ΔB c (i,t)-B x (i,t) (96)

[0160] η evs (i,t)=B evs (i,t) / B c (i,t) (97)

[0161] Among them, B c (i,t) is the electricity cost of electric vehicle user i at time t, c ch (t) is the charging price at time t; B evs (i,t),η evs (i, t) are the benefits and benefit increase ratio of electric vehicle user i participating in demand response at time t;

[0162] Comprehensively examine the average benefit B of all users participating in demand response at time t evm (t) and average return ratio η evm (t) is:

[0163]

[0164] In order to balance the needs and interests of all parties, the grid-side control target power is used as the dispatch constraint, and the maximization of the aggregator's profit increase ratio and the user's average revenue increase ratio is used as the comprehensive optimization goal, namely:

[0165] maxF(t)=β1η EVA (t)+β2η evm (t) (99)

[0166] Among them, β1 and β2 are the weight coefficients of the aggregator's profit growth ratio and the user's average revenue ratio respectively.

[0167] In this embodiment, the conventional load of a commercial area is used as the base load, and the electric vehicle predicted load is determined by sampling the probability distribution of the charging time and initial power of the electric vehicles in the commercial area, such as Figure 2 The control target power is taken as the power value when the voltage level is 10kV, the conductor current carrying capacity is 381A, and the power factor is 0.95; the time-of-use electricity price for electric vehicle charging is consistent with that of a charging station in Guangzhou, as shown in Table 1. The total number of electric vehicles arriving at the electric vehicle aggregator every day is N v 600 vehicles, P cN Take 60kW, C0 take 60kW·h, η take 0.95; S dm Take 1; the length of stay of electric vehicles obeys the normal distribution of N(2,0.5); S M Take 0.9; S zd Obey the uniform distribution of U(0.6, 0.9); b EVA Take 2.5 yuan / (kW·h), c0 for category B and C users are 1.5 yuan / (kW·h) and 2.5 yuan / (kW·h) respectively, c m Take 3 yuan / (kW·h) and 4 yuan / (kW·h) respectively; a x 、b x Take 0.005 yuan / (kW·h) respectively 2 and 1.5 yuan / (kW·h); Δt is 15 min; according to the size of the benefit ratio, the ratio of β1 and β2 is set to 1:4, and the sum is 1.

[0168] Table 1

[0169]

[0170] When the proportion of contracted users is 30%, the total load curves before and after optimization when ordinary users participate in the response are as follows: Figure 3a As shown, the EV load curve is as follows Figure 3b As shown, the number of responding vehicles and average power are as follows Figure 3c The optimization results are shown in Table 2.

[0171] As can be seen, starting from the 41st period, as the predicted load exceeded the control target, the EVA determined a scheduling plan based on the scheduling demand and the scheduling potential of each EV type and dispatched the corresponding EVs. After scheduling, the EV charging load peak reduction rate reached 8.35%. Due to the large base load, the total load peak reduction rate dropped to 7.68%. When the load suddenly increased, the number of EVs participating in the response increased rapidly, while the average charging power of the participating EVs was regulated and rapidly decreased. Near the original load peak period, the power reduction was also close to the peak value, verifying the effectiveness of the real-time scheduling strategy proposed in this paper.

[0172] As shown in Table 2, during the entire dispatch process, the total number of response periods was 26, with an average of 17 EVs participating in each response period. The average response power of these vehicles was 25 kW, and the average delayed charging time was 18.9 minutes. The total revenue of the aggregator was 48,871 yuan, with a revenue growth rate of 6.84. The average revenue of the EV users who participated in the response was 46.71 yuan, with an average revenue growth rate of 1.66. While trying to keep the total load within the target power control on the grid side, the interests of both the aggregator and the users can be taken into account.

[0173] Table 2

[0174]

[0175] Example 2:

[0176] In this embodiment, when the proportion of contracted users is 30%, the total load curve of whether ordinary users participate in the response or not is as follows: Figure 4 The optimization results are shown in Table 3.

[0177] Table 3

[0178]

[0179] As can be seen, compared to not participating in the response, the total load peak reduction rate increases from 2.27% to 7.68% when ordinary users participate. This is because when ordinary users participate in the response, the dispatchable potential of each time period is greatly increased, resulting in a significant reduction in load peaks. When ordinary users do not participate in the response, too few EVs can be dispatched to meet grid demand. The grid will reduce subsidies to aggregators, which will reduce their revenue and revenue growth rate. This will also reduce the average revenue and average revenue growth rate of participating EVs. Therefore, it is necessary to include ordinary users in the DR scope. This will not only better meet grid demand, but also increase the revenue of aggregators and users, achieving a win-win situation for all three parties.

[0180] Example 3:

[0181] In this embodiment, when the proportion of contracted users is different, the optimization effect of real-time scheduling is also different. Figure 5 The load curves of ordinary users participating in the response are shown when the proportion of contracted users is 15%, 30% and 45% respectively. The optimization results are shown in Table 4.

[0182] Table 4

[0183]

[0184] Obviously, the above embodiments of the present invention are merely examples for the purpose of clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. Those skilled in the art will appreciate that other variations or modifications can be made based on the above description. It is not necessary and impossible to enumerate all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the claims of the present invention.

Claims

1. A real-time dispatching method for electric vehicles taking into account predicted load and user demand, characterized in that: The steps include: S1. EV users are classified based on whether they have signed a contract with an EV aggregator and their charging needs, and charging control models are established accordingly. Taking lithium batteries as the object, the battery self-discharge process is ignored and it is approximately assumed that the battery maintains a constant charging power within each optimization period. The charging model of EVs is established as follows: Among them, S ne is the charging power requirement of the electric vehicle; S0 is the initial state of charge when the electric vehicle is connected; S ex is the expected state of charge of the electric vehicle when it leaves; C0 is the battery capacity; S(t) is the state of charge of the electric vehicle at time t; η is the charging efficiency; P c (t) is the charging power of the electric vehicle at time t; Δt is the time interval; P cN is the rated charging power of the electric vehicle; t ar With t ex are the arrival time and estimated departure time of the electric vehicle, respectively; Determine the power boundary of the electric vehicle as follows: T cmin =C0(S ex -S0) / (ηP cN ) (2) t ml =t ar +T cmin (3) t ex =t ar +T tl (5) t mc =t ex -T cmin (6) Among them, T cmin The shortest charging time for electric vehicles; t ml is the fastest departure time of the electric vehicle; S max (t) is the upper limit of the electric vehicle's power at time t in the station; T tl is the length of time the electric vehicle stays in the station; t mc The latest time to start charging the electric vehicle; S min (t) is the lower bound of the electric vehicle’s charge at time t when it is in the station; Assume that the contracted user will actively participate in demand response when there is no urgent travel demand, and set a scheduling power difference limit S dm As a restriction condition for the aggregator to freely dispatch contracted users, it is used to prevent contracted users without urgent travel needs from being in a dispatched state and unable to fully charge. That is, when the difference between the current power of a contracted user and the power charged at the rated power after entering the station exceeds the dispatch power difference limit S dm , then the electric vehicle will no longer participate in the dispatch; the contracted users without urgent travel needs are recorded as Class A users. The charging control model for Class A users is as follows: Where q(i) is the category of electric vehicle i; d(i) is the dispatch priority of electric vehicle i, d(i) = 1 indicates the highest dispatch priority, d(i) = 2 and d(i) = 3 indicate the dispatch priority is lower in turn; S ex (i) is the expected power of electric vehicle i; S M The maximum charging capacity of the electric vehicle; P min (i,t) and P max (i, t) are the lower and upper limits of the charging power of electric vehicle i at time t; s(i, t) is the dispatchable state of electric vehicle i at time t, s(i, t) = 0 means it is not dispatchable, and s(i, t) = 1 means it is dispatchable. Contracted users with urgent travel needs are always in the non-dispatchable state, as follows: Among them, S b (i, t) is the charging power of electric vehicle i at rated charging power P cN Time from arrival to destination t ar The standard power charged to time t; S d (i, t) is the dispatch power difference of electric vehicle i at time t; S(i, t) is the power of electric vehicle i at time t; S0(i) is the initial state of charge when electric vehicle i is connected; For users who have not signed contracts with aggregators, that is, ordinary users, they often do not choose long-term contracts due to the high flexibility of their daily travel needs. If ordinary users have low travel needs when the aggregator issues a demand response, and hope to participate in demand response in exchange for certain benefits, they can participate in demand response by signing a temporary contract and accept the aggregator's dispatch. Therefore, for ordinary users who are willing to participate in demand response, if their stay time at the charging station exceeds the minimum charging time, they can participate in demand response without reducing their desired charging power. Such users are recorded as Class B users, and their scheduling priority is second only to Class A users. When Class A users cannot meet the grid dispatch requirements, the aggregator needs to use the compensation price for Class B users to guide Class B users to participate in demand response. For ordinary users whose stay time at the charging station is less than the minimum charging time, they need to participate in demand response with the minimum charging power as their charging demand. They are recorded as Class C users, and their scheduling priority is lower than that of Class A and Class B users. When Class A and Class B users cannot meet the grid dispatch requirements, the aggregator needs to pay a higher compensation price than the compensation price for Class B users to guide Class C users to participate in demand response. The charging control model for ordinary users is as follows: Among them, S zd (i) The minimum charging power provided to electric vehicle i, which is lower than the maximum charging power S of electric vehicle M ;S ex (i) is the expected state of charge of electric vehicle i when it leaves; T cmin (i) is the shortest charging time of electric vehicle i; T tl (i) is the length of time that electric vehicle i stays at the station; S min (i, t) is the lower bound of the amount of electricity consumed by electric vehicle i at the station at time t, S max (i, t) is the upper bound of the amount of electricity that electric vehicle i can consume at the station at time t; S2. Establish a relationship model between the response probability of non-contracted users and the compensation electricity price; S3. Establish a subsidy mechanism for the grid to subsidize aggregators and for aggregators to subsidize electric vehicle users participating in demand response; S4. Calculate the response potential of various types of electric vehicle users, determine the grid-side dispatching demand for each time period based on the predicted load, and formulate a real-time dispatching plan for each time period based on the grid-side dispatching demand and the potential of electric vehicle users; S5. Establish a real-time optimization model, take the grid-side control target power as the constraint, take the profit increase ratio of electric vehicle aggregators and the average revenue increase ratio of electric vehicle users as the comprehensive objectives, and use the particle swarm algorithm to solve the electric vehicle charging power.

2. The real-time dispatching method for electric vehicles taking into account predicted load and user demand according to claim 1, characterized in that: In step S2, a relationship model between the response probability of non-contracted users and the compensation electricity price is established, as follows: When the dispatching potential of contracted users cannot meet the grid demand, it is necessary to formulate a reasonable compensation price to guide ordinary users to participate in the response. Since the higher the compensation price, the higher the benefits of users participating in demand response and the greater the response probability, a direct proportional function is used to establish the relationship between the response probability of non-contracted users, i.e. ordinary users, and the compensation price, as follows: μ=1 / (c m -c0) (13) Among them, p x (t) is the response probability of ordinary users at time t; c(t) is the compensation price at time t; μ is the response probability of ordinary users increased by increasing the unit compensation price; c0 and c m They are the minimum and maximum compensation electricity prices that aggregators provide to ordinary users respectively.

3. The real-time dispatching method for electric vehicles taking into account predicted load and user demand according to claim 2, characterized in that: Step S3 includes the following steps: S3.

1. Establish a compensation mechanism for the power grid to aggregators. When aggregators participate in demand response, the power grid needs to compensate the aggregators based on the degree of load reduction, as follows: B g (t)=W y (t)·b EVA ·a (14) D f (t)=W(t) / P ne (t) (16) Among them, B g (t) is the compensation paid by the power grid to the aggregator at time t; W y (t) is the effective response power of the aggregator at time t; b EVA is the subsidy standard for participating in the response, ranging from 0 to 5 (yuan / kW·h); a is the response coefficient, which is 3 under the real-time peak-shaving demand response; D f (t) is the response completion of the aggregator at time t; P ne (t) is the scheduling demand at time t; N s (t) is the sum of the number of electric vehicles that have arrived and the number of electric vehicles expected to arrive participating in demand response at time t; P(i,t) is the charging power of electric vehicle i at time t; S3.

2. Establish a compensation mechanism for electric vehicle users by aggregators, subsidizing electric vehicle users based on the reduced power consumption of electric vehicles, as follows: B ev (i,t)=c(t)·W(i,t) (19) W(i,t)=Δt·(P cN -P(i,t)) (20) Among them, B EV (t) is the total subsidy cost of the aggregator to all electric vehicles at time t; B ev (i,t) is the compensation fee paid by the aggregator to electric vehicle i at time t; W(i,t) is the response power of electric vehicle i at time t.

4. The method for real-time dispatch of electric vehicles taking into account predicted load and user demand according to any one of claims 1 to 3, characterized in that: Step S4 includes the following steps: S4.

1. The dispatch potential of each type of electric vehicle and the dispatch demand in each time period are calculated as follows: P ne (t)=P pev (t)+P pb (t)-P aim (21) P pev (t)=P cN (N n (t)+N p (t)) (22) Among them, P ne (t) is the scheduling demand at time t, P pev (t) is the electric vehicle load predicted to be connected before the next moment at moment t, P pb (t) is the normal load predicted at time t, P aim is the control target power on the grid side; N n (t) is the number of electric vehicles at the charging station at time t; N p (t) is the number of electric vehicles expected to arrive at the charging station before the next moment; P cap (q,t) is the dispatch potential of q-type electric vehicles at time t; N s (q,t) is the number of electric vehicles of type q at time t, q = A, B or C; S4.

2. Develop real-time dispatch plans for each period based on the dispatch demand of the grid side and user potential. When there is a dispatch demand P in a certain period, ne (t), various types of electric vehicles are dispatched in descending order of dispatch priority. When the dispatch potential of high-priority users cannot meet the dispatch demand of the power grid, the electric vehicles of the lower level are dispatched.

5. The real-time dispatching method for electric vehicles taking into account predicted load and user demand according to claim 4 is characterized in that: In step S4.2, the specific scheduling scheme is as follows: 1) When P ne (t) <P cap (A, t), some Class A electric vehicles are dispatched within time t; 2) When P cap (A,t)≤P ne (t) <P cap (A,t)+P cap (B, t), all Class A electric vehicles and some Class B electric vehicles are dispatched within time t; 3) When P cap (A,t)+P cap (B,t)≤P ne (t) <P cap (A,t)+P cap (B,t)+P cap (C,t), dispatches all A and B class electric vehicles and some C class electric vehicles at time t; 4) When P cap (A,t)+P cap (B,t)+P cap (C,t)≤P ne At time (t), all electric vehicles are dispatched within time t.

6. The method for real-time dispatch of electric vehicles taking into account predicted load and user demand according to claim 4, characterized in that: S5 includes the following steps: S5.

1. Constraints for constructing a real-time optimization model based on the grid-side control target power; S5.

2. The comprehensive objective of constructing a real-time optimization model is to use the profit growth ratio of electric vehicle aggregators and the average revenue growth ratio of electric vehicle users.

7. The method for real-time dispatch of electric vehicles taking into account predicted load and user demand according to claim 6, characterized in that: In step S5.1, the real-time optimization model uses the grid-side control target power as a constraint, as follows: Among them, P(i,t) is the charging power of electric vehicle i at time t; P u (t) is the sum of the non-dispatchable loads at time t, including the electric vehicle load and conventional load in the non-dispatchable state; P aim is the control target power N on the grid side s (t) is the sum of the number of electric vehicles that have arrived and the number of electric vehicles expected to arrive participating in demand response at time t.

8. The real-time dispatching method for electric vehicles taking into account predicted load and user demand according to claim 7, characterized in that: In step S5.2, the details are as follows: The aggregator’s costs include the loss of service fee revenue ΔB at time t ser (t) and compensation to users B EV (t), the benefit is the compensation fee B from the power grid company at time t g (t), that is: B EVA (t)=B g (t)-ΔB ser (t)-B EV (t) (26) η EVA (t)=B EVA (t) / B ser (t) (27) Among them, B ser (t) and B' ser (t) is the service fee income expected before and after optimization at time t; P(t) is the total charging load expected after optimization at time t; B EVA (t) is the expected profit of the aggregator after optimization at time t; c ser (t) is the charging service price at time t; η EVA (t) is the expected profit growth ratio of the aggregator after optimization at time t; For electric vehicle users, their costs must include response costs. The user response cost has a monotonic and concave characteristic with respect to power reduction, so it is represented by a quadratic function: B x (i,t)=a x [W(i,t)] 2 +b x W(i,t) (28) Among them, B x (i, t) is the response cost of electric vehicle user i at time t; W(i, t) is the reduced power consumption of electric vehicle user i at time t; a x and b x are coefficients, all of which are constants greater than 0; User benefits include the reduced electricity cost ΔB of electric vehicle user i at time t c (i, t) and the compensation fee B paid by the aggregator to electric vehicle user i at time t ev (i,t), that is: B evs (i,t)=B ev (i,t)+ΔB c (i,t)-B x (i,t) (30) η evs (i,t)=B evs (i,t) / B c (i,t) (31) Among them, B c (i,t) is the electricity cost of electric vehicle user i at time t, c ch (t) is the charging price at time t; B evs (i,t),η evs (i, t) are the benefits and benefit increase ratio of electric vehicle user i participating in demand response at time t; Comprehensively examine the average benefit B of all users participating in demand response at time t evm (t) and average return ratio η evm (t) is: In order to balance the needs and interests of all parties, the grid-side control target power is used as the dispatch constraint, and the maximization of the aggregator's profit increase ratio and the user's average revenue increase ratio is used as the comprehensive optimization goal, namely: maxF(t)=β1η EVA (t)+β2η evm (t) (33) Among them, β1 and β2 are the weight coefficients of the aggregator's profit growth ratio and the user's average revenue ratio respectively.

Images