A method and system for electric vehicle aggregation and bidding for peak regulation ancillary services

Abstract

The application relates to a kind of electric vehicle aggregation and bidding method and system for peak-shaving auxiliary service, comprising the following steps: based on electric vehicle operating parameter data, the dynamic model of single electric vehicle is established;Based on the dynamic model of single electric vehicle, the dynamic model of electric vehicle cluster aggregation is obtained by approximate optimization solution through aggregation algorithm;Based on the dynamic model of electric vehicle cluster aggregation, the double-layer strategy model of electric vehicle aggregator participating in peak-shaving market is established;The double-layer strategy model of electric vehicle aggregator participating in peak-shaving market is converted into a single-layer mathematical optimization function containing equality constraints for solving by Karush-Kuhn-Tucker condition and strong duality principle, and the purchase plan of participating in power market and the amount and price of peak-shaving auxiliary service market are obtained.The method provided by the application can quickly aggregate large-scale, parameter-heterogeneous electric vehicle cluster, and the aggregator can not only optimize its income in the power market, but also effectively participate in power grid peak-shaving auxiliary service and play its regulating role.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of demand side resource management of power systems, and particularly relates to an electric vehicle aggregation and bidding method and system for peak shaving auxiliary services. BACKGROUND

[0002] In recent years, the proportion of new energy power generation represented by wind power and photovoltaic power has been increasing year by year. Although these new energies have the advantages of being renewable and low carbon emission, their intermittency and volatility have brought serious challenges to the stable operation of the power grid. Especially during the peak load period of the power grid, the traditional peak shaving method relies on the expansion of peak power supply and power transmission channels, which not only significantly increases the operation cost of the power grid, but also fails to fully utilize the flexibility of the power grid resources.

[0003] In the demand side management of the power grid, electric vehicles are considered as a key resource for alleviating the volatility of the power grid and improving the flexibility of the system due to their potential large-scale energy storage capacity and flexible load regulation characteristics. It is predicted that the number of electric vehicles will reach 83 million, the equivalent energy storage capacity will reach 5 TW·h, the demand for electric charging will account for 6% to 7% of the total social electricity consumption, and the maximum charging load will account for 11% to 12% of the power grid load. However, due to the wide distribution of electric vehicles, the heterogeneous parameters and the limited individual regulation capacity, a single electric vehicle cannot meet the requirements of directly participating in the power market.

[0004] To address the above challenges, the introduction of electric vehicle aggregators (Aggregators) is crucial. By aggregating a large number of dispersed demand side electric vehicles into a whole, the requirement for regulation capacity is met, and the electric vehicles can effectively participate in the power market and interact with the power grid. The electric vehicle aggregator is responsible for directly regulating these resources to achieve the goals of peak shaving, new energy consumption and peak clipping of the power system.

[0005] The existing aggregation method has the challenges of high computational complexity and insufficient approximation accuracy when dealing with the parameter heterogeneity and large-scale aggregation of electric vehicles. In particular, when facing the non-full-dimensional characteristics of the power domain of electric vehicles, the traditional Minkowski and (M-sum) solving method is difficult to achieve efficient and accurate aggregation. SUMMARY

[0006] The application aims at the problem that power domain polyhedron is not full-dimensional due to parameter heterogeneity of electric vehicles, and it is difficult to solve approximately, and provides an electric vehicle aggregation and bidding method for peak regulation auxiliary service, which improves the calculation efficiency and approximation accuracy while solving the non-full-dimensional approximation problem through an improved approximate solution strategy, thereby releasing more electric vehicle regulation flexibility and being able to quickly aggregate large-scale, parameter-heterogeneous electric vehicle clusters. Further considering the electricity purchase cost, a bidding strategy of the electric vehicle aggregator in the auxiliary service market is designed, and the aggregator can not only optimize its income in the electricity market, but also effectively participate in the grid peak regulation auxiliary service and play its regulation role.

[0007] The application solves the above technical problems in the following scheme: an electric vehicle aggregation and bidding method for peak regulation auxiliary service, comprising the following steps:

[0008] Based on the electric vehicle operation parameter data, a dynamic model of a single electric vehicle is established;

[0009] Based on the dynamic model of a single electric vehicle, an aggregated dynamic model of an electric vehicle cluster is obtained by approximate optimization through an aggregation algorithm;

[0010] Based on the aggregated dynamic model of the electric vehicle cluster, a double-layer strategy model of the electric vehicle aggregator participating in the peak regulation market is established;

[0011] The double-layer strategy model of the electric vehicle aggregator participating in the peak regulation market is converted into a single-layer mathematical optimization function with equality constraints through the Karush-Kuhn-Tucker condition and the strong duality principle for solving, and the electricity purchase plan for participating in the electricity market and the bid quantity and price for the peak regulation auxiliary service market are obtained.

[0012] Preferably, the dynamic model of a single electric vehicle is established based on the electric vehicle operation parameter data, comprising:

[0013] Based on the electric vehicle operation parameter data, including the power, state of charge and charge / discharge time limit of the electric vehicle, a dynamic model of a single electric vehicle is established, specifically as follows:

[0014]

[0015] wherein E i (t) is the energy state of the i-th electric vehicle at time t, P i ch (t) / P i dis (t) is the charge / discharge power of the i-th electric vehicle at time t, is the upper limit of the charge / discharge power of the i-th electric vehicle at time t, is the lower / upper bound of the state of charge of the ith electric vehicle at time t, α i is the energy dissipation rate of the ith electric vehicle, Δt is the sampling time interval, t i,a i,l is the arrival / departure time of the electric vehicle at time t, is the expected state of charge of the EVi when it leaves.

[0016] Preferably, the dynamic model based on a single electric vehicle is approximated and optimized by an aggregation algorithm to obtain a dynamic model of the aggregation of the electric vehicle cluster, specifically including:

[0017] The dynamic model of a single electric vehicle is converted into a polyhedral half-space form, specifically as follows:

[0018]

[0019] The power constraint of the electric vehicle is converted into a closed region representation in a multi-dimensional linear space, also known as the power feasible region, which represents the power set of the electric vehicle within a certain time, facilitating subsequent Minkowski sum and aggregation solving.

[0020] In the formula, P i represents the D time node power of the ith electric vehicle, is the set of power curves of the ith electric vehicle, representing the power range of the ith electric vehicle, that is, the power feasible region of EVi. Set P i is the actual input / output power, when P i ≥ 0, it is the charging state, when P i ≤ 0, it is the discharging state, The content after the symbol is a further explanation of the content before the symbol.

[0021] Half-space representation A i of the dynamic model of a single electric vehicle i ≤ b i , A i , b i are as follows:

[0022]

[0023] where E i (0) is the electric energy state of the ith electric vehicle at time t, D ≥ t i,l -t i,a , D is the number of aggregation times aggregated by the aggregator, and the specific matrix form is as follows:

[0024] ​

[0025]

[0026] Based on the dynamic model transformation of individual electric vehicles and its polytope in half-space form, the aggregated dynamic model of the cluster of electric vehicles is approximately solved by Minkowski sum and M-Sum, as follows:

[0027] The aggregated dynamic model of the cluster of electric vehicles is expressed in Minkowski sum and M-Sum of N electric vehicle dynamic models as follows:

[0028]

[0029] where, is the cluster of N electric vehicles, is the aggregated dynamic model of the cluster of electric vehicles, whose result is still in the form of linear constraints Ax≤b. denotes direct sum.

[0030] Directly calculating the exact M-Sum of N half-space representation polytopes is an NP-hard problem, so the approximate solution method is generally used to transform the NP-hard problem into the problem of arithmetic accumulation, reducing the computational complexity. In view of the problem that the existing approximate algorithm cannot aggregate the electric vehicles with heterogeneous arrival / departure times t i,a / t i,l , a new approximate solution method is used to calculate the NP-hard problem of N half-spaces, as follows:

[0031] The power feasible region of N electric vehicles is approximately expressed in the form of having the same matrix as The aggregated power feasible region of the cluster of electric vehicles can be approximately expressed as , transforming the complex calculation of the exact M-sum in the aggregation of electric vehicles into the calculation of the arithmetic sum of .

[0032] The Hausdorff distance is used to measure the approximation quality, and then the external and internal approximation problems can be described as follows:

[0033] or

[0034] where, is the external approximation of the power feasible region of EV i, is the internal approximation of the power feasible region of EV i.

[0035] External approximation constraints mean that the external approximate power feasible region includes the entire set of the operable power curves of the electric vehicle and a small portion of the inoperable power curves. Internal approximation constraints reduce the power feasible region of the electric vehicle to include a portion of the operable power curves.

[0036] By utilizing the trigonometric inequality of the Hausdorff distance and the Lipschitz continuity, the objective function is transformed into a convex function of L2 norm form, as follows:

[0037] Lemma 1: There are two power feasible regions for electric vehicles and If ||·|| p ,||·|| q Let be two arbitrary norms and d H (·,·) is derived from the norm ||·|| p The Hausdorff distance is measured, and therefore the Hausdorff distance across the power feasible region of the two electric vehicles satisfies:

[0038]

[0039] in,

[0040]

[0041] According to Lemma 1 and the Hausdorff distance triangle inequality;

[0042] right Simplify:

[0043] Construct a non-empty set:

[0044] According to the Hausdorff distance triangle inequality, we can obtain:

[0045] because As a fixed value, L(A0) is the Lipschitz constant of matrix A0, therefore the objective function... A convex function that can be equivalently converted into a 2-norm form

[0046] Based on Farkas' lemma, the set inclusion relation constraints in the approximate optimization problem are transformed into linear and nonlinear constraints. For the nonlinear constraints, the basic polyhedral scaling algorithm is used for solution, as follows:

[0047] The constraints for the external approximation are linear constraints, as follows:

[0048]

[0049] The inner approximation condition is a nonlinear constraint, as follows:

[0050]

[0051] For cases where the inner approximation condition is a nonlinear constraint, the outer approximation polyhedron is solved. Treating it as a basic polyhedron and solving it using the basic polyhedron scaling algorithm, the internal approximation problem is simplified to:

[0052]

[0053] Among them, s i =1 / φ i ,r i =-s i ψ i ;

[0054] use An accurate dynamic model that approximates electric vehicles The aggregation dynamic model of the electric vehicle cluster is as follows: The symbol is a logical sign indicating that two conditions are equivalent.

[0055]

[0056] Preferably, the dynamic model based on electric vehicle cluster aggregation establishes a two-layer strategy model for electric vehicle aggregators to participate in the peak-shaving market, including:

[0057] Based on the dynamic model of electric vehicle cluster aggregation, a model for electric vehicle aggregators to participate in the electricity market that minimizes charging costs within the feasible aggregation region of electric vehicle clusters is established, as shown in the following expression:

[0058]

[0059] This indicates that within the feasible aggregation domain of the electric vehicle cluster, the charging cost for the electric vehicle aggregator is minimized.

[0060] in, Let λ be the charging power of the electric vehicle aggregator, and λ be the decision variable. EM (t) publishes time-of-use pricing for the electricity market, when When this occurs, it means that the aggregator charges the grid in reverse to receive payment;

[0061] Based on the model of electric vehicle aggregators participating in the electricity market, a high-level strategy model for electric vehicle aggregators to participate in the peak-shaving ancillary service market is established, which minimizes the overall cost of electric vehicle aggregators. The expression is as follows:

[0062]

[0063] The cost of the electric vehicle aggregator includes two parts, λ EM (t)P agg (t) is the purchase cost of the aggregator participating in the electricity market, is the optimal power curve of the electric vehicle aggregator participating in the electricity market only, calculated according to the calculation, is the power adjustment amount of the electric vehicle aggregator, is the income of the electric vehicle aggregator participating in the peak shaving auxiliary service market, the peak shaving auxiliary service market settles the income according to the adjustment amount of the peak shaving participant, λ(t) is the clearing price of the peak shaving auxiliary service market; λ(t)>0 indicates that the peak shaving demand is positive, and λ(t)<0 indicates that the peak shaving demand is positive.

[0064] The electric vehicle aggregator and the large battery participate in the peak shaving auxiliary service market, and the lower strategy model of the electric vehicle aggregator participating in the peak shaving market is established to minimize the overall peak shaving cost under the premise of meeting the peak shaving demand, and the expression is as follows:

[0065] Objective function:

[0066] The objective function is to minimize the overall peak shaving cost, including the peak shaving cost settled to the electric vehicle and the calling cost λ o (t)P o (t) of the large peak shaving device;

[0067] Constraint conditions: including the reported peak shaving range constraint of the electric vehicle aggregator, according to A0P agg (t)<=b agg calculation; is the maximum power that can be reached at each time node; is the minimum power that can be reached at each time node, is the up and down range of the optimal power curve difference of participating in the electricity market only, and the corresponding dual variable is as follows:

[0068]

[0069] The adjustable range constraint of the large peak shaving device, and the corresponding dual variable is as follows:

[0070]

[0071] The peak shaving supply and demand balance constraint, and the corresponding dual variable λ(t) is the clearing price of the peak shaving auxiliary service market, as follows:

[0072]

[0073] Preferably, the double-layer strategy model of the electric vehicle aggregator participating in the peak shaving market is converted into a single-layer mathematical optimization function with equality constraints by the Karush-Kuhn-Tucker condition and the strong duality principle, and the purchase plan of the electric power market and the bid amount and bid price of the peak shaving auxiliary service market are obtained, which specifically includes:

[0074] The Lagrange function is constructed for the lower-layer strategy model of the electric vehicle aggregator participating in the peak shaving market, and the objective function and the constraint condition of the lower-layer model are combined with the Lagrange multiplier, as follows:

[0075]

[0076] The KKT condition is applied to the lower-layer strategy model of the electric vehicle aggregator participating in the peak shaving market, as follows:

[0077]

[0078] The complementary slackness condition in the KKT condition is as follows:

[0079]

[0080] According to the strong duality principle, the optimal solution of the lower-layer strategy model of the electric vehicle aggregator participating in the peak shaving market is converted into the constraint condition of the upper-layer strategy model of the electric vehicle aggregator participating in the peak shaving market, and the double-layer model is converted into a single-layer model, as follows:

[0081]

[0082] According to the KKT condition and the strong duality principle, the double-layer strategy model of the electric vehicle aggregator participating in the peak shaving auxiliary service market is converted into a single-layer mathematical optimization problem with equality constraints, as follows:

[0083] The objective function can be obtained according to the complementary slackness condition:

[0084]

[0085] According to the KKT condition, the following can be obtained:

[0086]

[0087] The above KKT condition result is brought into the complementary slackness condition result, and the following can be obtained:

[0088]

[0089] The strong duality principle is further brought in, and the following can be obtained:

[0090]

[0091] Therefore, the objective function of the single-layer model of the electric vehicle aggregator participating in the peak regulation market is finally obtained as follows:

[0092]

[0093] The constraint conditions include the adjustable range constraint of the electric vehicle aggregator, the reported peak regulation range constraint, the adjustable range constraint of the peak regulation device, and the KKT condition constraint, as follows:

[0094] A0P agg (t) <= b agg

[0095]

[0096] The nonlinear function sign(P o (t)) and the complementary relaxation condition containing a bilinear term are linearized, as follows:

[0097] Linearization of the sign function sign(P o (t)) is as follows:

[0098]

[0099] Linearization of the complementary relaxation condition is as follows:

[0100]

[0101] By solving the double-layer strategy model of the electric vehicle aggregator participating in the peak regulation market, a mixed integer linear programming problem is converted, in which the continuous decision variables are P agg (t), P o (t), alpha agg (t), lambda(t), The decision variables are The purchase plan of participating in the power market and the reported amount and the price

[0102] The application also provides an electric vehicle aggregation and bidding system for peak regulation auxiliary services, comprising:

[0103] A dynamic model construction module of the electric vehicle is used to establish a dynamic model of a single electric vehicle based on electric vehicle operation parameter data;

[0104] A dynamic model construction module of the electric vehicle cluster aggregation is used to obtain a dynamic model of the electric vehicle cluster aggregation through approximate optimization solution by an aggregation algorithm based on the dynamic model of a single electric vehicle;

[0105] The double-layer strategy model construction module for the aggregator participating in the peak regulation market is configured to construct a double-layer strategy model for the aggregator of electric vehicles participating in the peak regulation market based on a dynamic model of the aggregator of the electric vehicle cluster.

[0106] The double-layer strategy model solution module for the aggregator participating in the peak regulation market is configured to convert the double-layer strategy model for the aggregator of electric vehicles participating in the peak regulation market into a single-layer mathematical optimization function with equality constraints through Karush-Kuhn-Tucker conditions and strong duality principle to solve the function to obtain the power purchase plan of the aggregator participating in the power market and the bid quantity and price of the aggregator participating in the peak regulation auxiliary service market.

[0107] Preferably, the dynamic model of the single electric vehicle is established based on electric vehicle operating parameter data, and includes the following steps.

[0108] The dynamic model of the single electric vehicle is established based on electric vehicle operating parameter data, including power, state of charge and charging / discharging time limit of the electric vehicle, and specifically includes the following steps.

[0109]

[0110] wherein, E i (t) is the energy state of the ith electric vehicle at time t, P i ch (t) / P i dis (t) is the charging / discharging power of the ith electric vehicle at time t, is the upper limit of the charging / discharging power of the ith electric vehicle at time t, is the lower / upper limit of the energy state of the ith electric vehicle at time t, α i is the energy dissipation rate of the ith electric vehicle, Δt is the sampling time interval, t i,a / t i,l is the arrival / departure time of the electric vehicle at time t, is the expected energy state of EVi when it departs.

[0111] Preferably, the dynamic model of the single electric vehicle is used to obtain the dynamic model of the electric vehicle cluster through an aggregation algorithm for approximate optimization, and specifically includes the following steps.

[0112] The dynamic model of the single electric vehicle is converted into a polyhedral half-space form, and specifically includes the following steps.

[0113]

[0114] wherein, P i represents the power of the ith electric vehicle at D time nodes, is the set of power profiles of the ith electric vehicle, which represents the range of power that the ith electric vehicle can operate at;

[0115]

[0116] where E i is the state of charge of the ith electric vehicle at time t, D≥t i,l -t i,a , D is the number of aggregation times of the aggregator, and the specific matrix form is as follows:

[0117]

[0118] Based on the conversion of the dynamic model of a single electric vehicle and the half-space form of its polyhedron, the aggregation dynamic model of the cluster of electric vehicles is approximately solved by using the Minkowski sum and M-Sum, and the specific process is as follows:

[0119] The aggregation dynamic model of the cluster of electric vehicles is expressed in the form of Minkowski sum and M-Sum of N electric vehicle dynamic models as follows:

[0120]

[0121] where is the cluster containing N electric vehicles, is the aggregation dynamic model of the cluster of electric vehicles;

[0122] The NP-hard problem of N half-spaces is calculated by using the approximate solution method, and the specific process is as follows:

[0123] The power feasible region of N electric vehicles is approximately expressed in the form of having the same matrix as The aggregation power feasible region of the cluster of electric vehicles can be approximately expressed as

[0124] The Hausdorff distance is used to measure the approximation quality, and then the external and internal approximation problems can be described as follows:

[0125] or

[0126] where is the external approximation of the power feasible region of EV i, is the internal approximation of the power feasible region of EV i;

[0127] ​By using the triangle inequality of Hausdorff distance and Lipschitz continuity, the objective function is transformed into a convex function in the form of two-norm as follows:

[0128] Simplify:

[0129] Construct a non-empty set:

[0130] According to the triangle inequality of Hausdorff distance, we have:

[0131] Because is a fixed value, and L(A0) is the Lipschitz constant of matrix A0, the objective function can be equivalently transformed into a convex function in the form of two-norm

[0132] According to Farkas' lemma, the set inclusion constraint in the approximate optimization problem is transformed into linear and nonlinear constraints. For the nonlinear constraint, the basic polytope scaling algorithm is used to solve it as follows:

[0133] The constraint condition of the outer approximation is linear, as follows:

[0134]

[0135] The inner approximation condition is nonlinear, as follows:

[0136]

[0137] For the case where the inner approximation condition is nonlinear, the outer approximation polytope is regarded as the basic polytope and the basic polytope scaling algorithm is used to solve it. The inner approximation problem is simplified as:

[0138]

[0139] where, i = 1 / φ i , r i = -s i ψ i ;

[0140] The approximate expression of the precise dynamic model of the electric vehicle is Then the aggregated dynamic model of the electric vehicle cluster is:

[0141]

[0142] ​​Preferably, the dynamic model based on the aggregation of electric vehicle clusters establishes a double-layer strategy model for electric vehicle aggregators participating in the peak shaving market, including:

[0143] The dynamic model based on the aggregation of electric vehicle clusters establishes an electric vehicle aggregator participating in the power market model for minimizing the charging cost of the electric vehicle aggregator within the aggregation feasible region of the electric vehicle cluster, expressed as follows:

[0144]

[0145] It represents the minimization of the charging cost of the electric vehicle aggregator within the aggregation feasible region of the electric vehicle cluster.

[0146] wherein, is the decision variable, is the charging power of the electric vehicle aggregator, λ EM (t) is the time-of-use price published by the power market, when , it represents the aggregator charging the grid in reverse to obtain compensation;

[0147] Based on the electric vehicle aggregator participating in the power market model, an upper strategy model for the electric vehicle aggregator participating in the peak shaving market is established, which simultaneously minimizes the overall cost of the electric vehicle aggregator and reports the amount of peak shaving, expressed as follows:

[0148]

[0149] The cost of the electric vehicle aggregator includes two parts, λ EM (t)P agg (t) is the purchase cost of the aggregator participating in the power market, is the optimal power curve of the electric vehicle aggregator calculated only participating in the power market, is the power adjustment amount of the electric vehicle aggregator, is the revenue of the electric vehicle aggregator participating in the peak shaving auxiliary service market, the peak shaving auxiliary service market settles the revenue according to the adjustment amount of the peak shaving participant, λ(t) is the clearing price of the peak shaving auxiliary service market; λ(t)>0 indicates that the peak shaving demand is positive, and λ(t)<0 indicates that the peak shaving demand is positive.

[0150] An upper strategy model for the electric vehicle aggregator participating in the peak shaving market is established, which minimizes the overall peak shaving cost under the premise of meeting the peak shaving demand, expressed as follows:

[0151] Objective function:

[0152] The objective function is to minimize the overall peak shaving cost, including the peak shaving cost settled to the electric vehicle and the calling cost of large-scale peak-shaving devices λ o (t) | P o (t) |;

[0153] Constraints:

[0154] Reporting peak-shaving range constraints of the electric vehicle aggregator, according to AOP agg (t) <= b agg Calculation;

[0155] Maximum power available for each time node; Minimum power available for each time node, Optimal power curve difference adjustment range for only participating in the electricity market, and its corresponding dual variable is As follows:

[0156]

[0157] Adjustable range constraints of large-scale peak-shaving devices, and its corresponding dual variable is As follows:

[0158]

[0159] Peak-shaving supply and demand balance constraints, and its corresponding dual variable λ(t) is the peak-shaving ancillary service market clearing price, as follows:

[0160]

[0161] Preferably, the double-layer strategy model of the electric vehicle aggregator participating in the peak-shaving market is converted into a single-layer mathematical optimization function with equality constraints by using the Karush-Kuhn-Tucker condition and the strong duality principle, to obtain the electricity purchase plan of participating in the electricity market and the reporting quantity and price of the peak-shaving ancillary service market, specifically including:

[0162] The Lagrangian function is constructed for the lower strategy model of the electric vehicle aggregator participating in the peak-shaving market, and the objective function and constraint conditions of the lower model are combined with the Lagrange multiplier, as follows:

[0163]

[0164] The KKT condition is applied to the lower strategy model of the electric vehicle aggregator participating in the peak-shaving market, as follows:

[0165]

[0166] The complementary slackness condition in the KKT condition is as follows:

[0167]

[0168] By strong duality principle, the optimal solution of the lower strategy model of the electric vehicle aggregator participating in the peak shaving market is transformed into the constraint condition of the upper strategy model of the electric vehicle aggregator participating in the peak shaving market, and the double-layer model is transformed into a single-layer model as follows:

[0169]

[0170] According to the KKT condition and strong duality principle, the double-layer strategy model of the electric vehicle aggregator participating in the peak shaving auxiliary service market is transformed into a single-layer mathematical optimization problem with equality constraints as follows:

[0171] The objective function can be obtained according to the complementary relaxation condition:

[0172]

[0173] According to the KKT condition, we have:

[0174]

[0175] The above KKT condition result is brought into the complementary relaxation condition result to obtain:

[0176]

[0177] Again, the strong duality principle is brought in to obtain:

[0178]

[0179] Therefore, the objective function of the single-layer model of the electric vehicle aggregator participating in the peak shaving market is finally obtained as:

[0180]

[0181] The constraint conditions include the adjustable range constraint of the electric vehicle aggregator, the reported peak shaving range constraint, the adjustable range constraint of the peak shaving device, and the KKT condition constraint, as shown below:

[0182] A0P agg (t)<=b agg

[0183]

[0184] The nonlinear function sign(P o (t)) and the complementary relaxation condition with bilinear terms are linearized respectively as follows:

[0185] Linearization of the sign function sign(P o (t)):

[0186]

[0187] Complementary slackness condition linearization:

[0188]

[0189] By solving the double-layer strategy model of the electric vehicle aggregator participating in the peak regulation market into a mixed integer linear programming problem, wherein the continuous decision variable is P agg (t), P o (t), alpha agg (t), lambda(t), The decision variable is Obtain the purchase plan of the electric power market and the quantity of the peak regulation auxiliary service market And the price

[0190] The application also provides a computer storage medium, the computer readable storage medium stores a computer program, the computer program is executed by a processor to realize the steps of the electric vehicle aggregation and bidding method for peak regulation auxiliary service as described above.

[0191] The beneficial effects of the application are: in view of the problem that the power domain polyhedron is not full-dimensional due to the heterogeneous parameters of electric vehicles, and it is difficult to approximate the solution, the improved approximate solution strategy is used to solve the non-full-dimensional approximation problem, improve the calculation efficiency and approximation accuracy, and release more electric vehicle regulation flexibility, so that large-scale, parameter-heterogeneous electric vehicle clusters can be quickly aggregated. Further, a strategic bidding mechanism is provided, considering the purchase cost, a bidding strategy of the electric vehicle aggregator in the auxiliary service market is designed. Through the strategy, the market competitiveness of the aggregator is improved, the income in the power market is optimized, and the aggregator can effectively participate in the peak regulation auxiliary service of the power grid and play its regulation role.

[0192] The above description is only a summary of the technical solutions of the application. In order to more clearly understand the technical means of the application and can be implemented according to the content of the specification, the following will be described in detail with the preferred embodiments of the application and the accompanying drawings. The specific embodiments of the application are described in detail by the following examples and the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0193] The drawings described herein are used to provide further understanding of the application, and form a part of the application. The schematic embodiments of the application and their descriptions are used to explain the application, and do not constitute an improper limitation on the application. In the drawings:

[0194] Figure 1 A flowchart of the electric vehicle aggregation and bidding method for peak regulation auxiliary service is provided for example 1.

[0195] Figure 2 A flow chart for the dynamic model of the electric vehicle cluster aggregation in embodiment 1 is constructed;

[0196] Figure 3 A module diagram of the electric vehicle aggregation and bidding system for peak regulation auxiliary service in embodiment 2. DETAILED DESCRIPTION

[0197] The principles and features of the present application are described below in conjunction with the accompanying drawings, and the examples are only used to explain the present application and not to limit the scope of the present application.

[0198] Embodiment 1

[0199] As shown in Figure 1 , the present application provides an electric vehicle aggregation and bidding method for peak regulation auxiliary service, comprising the following steps:

[0200] S1, based on the electric vehicle operating parameter data, a dynamic model of a single electric vehicle is established, specifically including:

[0201] Based on the electric vehicle operating parameter data, including the power, state of charge and charge / discharge time limit of the electric vehicle, a dynamic model of a single electric vehicle is established, specifically as follows:

[0202]

[0203] Wherein, E i (t) is the energy state of the i-th electric vehicle at time t, P i ch (t) / P i dis (t) is the charging / discharging power of the i-th electric vehicle at time t, is the upper limit of the charging / discharging power of the i-th electric vehicle at time t, is the lower / upper limit of the energy state of the i-th electric vehicle at time t, α i is the energy dissipation rate of the i-th electric vehicle, Δt is the sampling time interval, t i,a / t i,l is the arrival / departure time of the electric vehicle at time t, is the expected energy state when EVi leaves.

[0204] S2, as shown in Figure 2 , based on the dynamic model of a single electric vehicle, the dynamic model of the electric vehicle cluster aggregation is obtained by approximate optimization through aggregation algorithm;

[0205] The dynamic model of a single electric vehicle is converted into a polyhedral half-space form, specifically as follows:

[0206]

[0207] P i represents the D time node power of the i-th electric vehicle, is the set of the i-th electric vehicle power curve, which represents the i-th electric vehicle's available power range;

[0208]

[0209] E i is the i-th electric vehicle's energy state at time t, D≥t i,l -t i,a , D is the number of aggregation time points aggregated by the aggregator, and the specific matrix form is as follows:

[0210]

[0211] Based on the conversion of the dynamic model of a single electric vehicle and its polyhedral half-space form, the aggregation dynamic model of the cluster of electric vehicles is approximately solved by Minkowski sum and M-Sum, as follows:

[0212] The aggregation dynamic model of the cluster of electric vehicles is expressed as the Minkowski sum of the dynamic models of N electric vehicles as follows:

[0213]

[0214] wherein, is a cluster containing N electric vehicles, is the aggregation dynamic model of the cluster of electric vehicles;

[0215] The NP-hard problem of calculating N half-spaces is solved by an approximate method, as follows:

[0216] The power feasible region of N electric vehicles is approximately expressed as in the form of having the same matrix The aggregation power feasible region of the cluster of electric vehicles can be approximately expressed as

[0217] The average value of the parameter matrix A i within the cluster of electric vehicles is calculated to obtain the isomorphic parameter matrix A0;

[0218] The Hausdorff distance is used to measure the approximation quality, and the external and internal approximation problems can be described as follows, respectively:

[0219] ​ or

[0220] where, is the outer approximation of the power feasible region of EV i, is the inner approximation of the power feasible region of EV i;

[0221] By using the triangle inequality of Hausdorff distance and Lipschitz continuity, the objective function is transformed into a convex function in the form of two-norm as follows:

[0222] Simplifying

[0223] Constructing a non-empty set

[0224] According to the triangle inequality of Hausdorff distance, we have:

[0225] Because is a fixed value, L(A0) is the Lipschitz constant of matrix A0, thus the objective function can be equivalently transformed into a convex function in the form of two-norm

[0226] According to Farkas' lemma, the set inclusion constraint in the approximation optimization problem is transformed into linear constraints and nonlinear constraints; for the nonlinear constraints, the basic polytope scaling algorithm is used to solve them as follows:

[0227] The constraint condition of the outer approximation is a linear constraint as follows:

[0228]

[0229] The inner approximation condition is a nonlinear constraint as follows:

[0230]

[0231] Calculate the outer approximation parameter matrix of each electric vehicle

[0232] For the case where the inner approximation condition is a nonlinear constraint, the outer approximation polytope is regarded as the basic polytope and the basic polytope scaling algorithm is used to solve it. The inner approximation problem is simplified as:

[0233]

[0234] where, i = 1 / φ i r​i = -s i ψ i ;

[0235]

[0236] Calculate the internal approximation scaling factor φ of each electric vehicle i = 1 / s i ,ψ i = -r i / s i Then, the parameter matrix is calculated

[0237] The approximate expression of the accurate dynamic model of the electric vehicle is adopted The aggregation dynamic model of the electric vehicle cluster is:

[0238] S3, based on the aggregation dynamic model of the electric vehicle cluster, a double-layer strategy model of the electric vehicle aggregator participating in the peak shaving market is established, which specifically includes:

[0239] Based on the aggregation dynamic model of the electric vehicle cluster, the electric vehicle aggregator participating in the power market model is established, which minimizes the charging cost of the electric vehicle aggregator within the aggregation feasible region of the electric vehicle cluster, and the expression is as follows:

[0240]

[0241] Among them,

[0242] is the decision variable, is the charging power of the electric vehicle aggregator, λ EM (t) is the time-of-use price published by the power market, when , it means that the aggregator charges the grid in reverse to obtain compensation.

[0243] Based on the electric vehicle aggregator participating in the power market model, the upper strategy model of the electric vehicle aggregator participating in the peak shaving market is established, which simultaneously minimizes the overall cost of the electric vehicle aggregator, and the expression is as follows:

[0244]

[0245] The cost of the electric vehicle aggregator includes two parts, λ EM (t)P agg (t) is the purchase cost of the aggregator participating in the power market, is the optimal power curve of the electric vehicle aggregator only participating in the power market calculated, is the power adjustment amount of the electric vehicle aggregator​ To the benefit of the electric vehicle aggregator participating in the peak shaving auxiliary service market, the peak shaving auxiliary service market settles the benefit according to the adjustment amount of the peak shaving participant, and λ(t) is the clearing price of the peak shaving auxiliary service market; λ(t)>0 indicates that the peak shaving demand is positive, and λ(t)<0 indicates that the peak shaving demand is positive.

[0246] The electric vehicle aggregator and the large battery participate in the peak shaving auxiliary service market, and under the premise of meeting the peak shaving demand, a lower strategy model of the electric vehicle aggregator participating in the peak shaving market is established to minimize the overall peak shaving cost, and the expression is as follows:

[0247] a. Objective function:

[0248] The objective function is to minimize the overall peak shaving cost, including the peak shaving cost settled to the electric vehicle and the calling cost λ of the large peak shaving device o (t)|P o (t)|;

[0249] b. Constraint condition: including the reported peak shaving range constraint of the electric vehicle aggregator, calculated according to A0P agg (t)<=b agg ; the maximum power available for each time node; the minimum power available for each time node, the upper and lower adjustment range of the optimal power curve difference of only participating in the power market, and the corresponding dual variable is as follows:

[0250]

[0251] the adjustable range constraint of the large peak shaving device, and the corresponding dual variable is as follows:

[0252]

[0253] the peak shaving supply and demand balance constraint, and the corresponding dual variable λ(t) is the clearing price of the peak shaving auxiliary service market, as follows:

[0254]

[0255] S4, the double-layer strategy model of the electric vehicle aggregator participating in the peak shaving market is converted into a single-layer mathematical optimization function with equality constraints through the Karush-Kuhn-Tucker condition and the strong duality principle, and the purchase plan of participating in the power market and the reported amount and price of the peak shaving auxiliary service market are obtained, which specifically includes:

[0256] The Lagrangian function is constructed for the lower strategy model of the electric vehicle aggregator participating in the peak shaving market, and the objective function and constraint conditions of the lower model are combined with the Lagrange multiplier as follows:

[0257]

[0258] The KKT condition is applied to the lower strategy model of the electric vehicle aggregator participating in the peak shaving market as follows:

[0259]

[0260] The complementary slackness condition in the KKT condition is as follows:

[0261]

[0262] By the strong duality principle, the optimal solution of the lower strategy model of the electric vehicle aggregator participating in the peak shaving market is transformed into the constraint condition of the upper strategy model of the electric vehicle aggregator participating in the peak shaving market, and the double-layer model is transformed into a single-layer model as follows:

[0263]

[0264] According to the KKT condition and the strong duality principle, the double-layer strategy model of the electric vehicle aggregator participating in the peak shaving auxiliary service market is transformed into a single-layer mathematical optimization problem with equality constraints as follows:

[0265] The objective function can be obtained according to the complementary slackness condition as follows:

[0266]

[0267] According to the KKT condition, the following can be obtained:

[0268]

[0269] The above KKT condition result is brought into the complementary slackness condition result to obtain:

[0270]

[0271] Again, the strong duality principle can be obtained as follows:

[0272]

[0273] a. Therefore, the objective function of the single-layer model of the electric vehicle aggregator participating in the peak shaving market is finally obtained as follows:

[0274]

[0275] b. The constraints include the adjustable range constraints of the electric vehicle aggregator, the reported peak shaving range constraints, the adjustable range constraints of the peak shaving equipment, the KKT condition constraints, as shown below: A0P agg (t) <= b agg

[0276]

[0277] A0P agg (t) <= b agg

[0278]

[0279] The double-layer strategy model of the electric vehicle aggregator participating in the peak shaving market is converted into a mixed integer linear programming problem by solving, wherein the continuous decision variable is P agg (t), P o (t), a agg (t), l(t), The decision variable is The purchase plan of participating in the electricity market and the reported amount and the price of participating in the peak shaving auxiliary service market are obtained and the price

[0280] Embodiment 2

[0281] The embodiment provides an electric vehicle aggregation and bidding system for peak shaving auxiliary services, comprising:

[0282] A dynamic model construction module of an electric vehicle, configured to establish a dynamic model of a single electric vehicle based on electric vehicle operation parameter data;

[0283] A dynamic model construction module of an electric vehicle cluster aggregation, configured to obtain a dynamic model of an electric vehicle cluster aggregation by approximate optimization through an aggregation algorithm based on the dynamic model of a single electric vehicle;

[0284] A double-layer strategy model construction module of an aggregator participating in a peak shaving market, configured to establish a double-layer strategy model of an electric vehicle aggregator participating in a peak shaving market based on the dynamic model of the electric vehicle cluster aggregation;

[0285] A double-layer strategy model solving module of an aggregator participating in a peak shaving market, configured to convert the double-layer strategy model of the electric vehicle aggregator participating in the peak shaving market into a single-layer mathematical optimization function with equation constraints through Karush-Kuhn-Tucker conditions and strong duality principle to obtain a purchase plan of participating in an electricity market and a reported amount and a price of participating in a peak shaving auxiliary service market.

[0286] Embodiment 3

[0287] The application further provides a computer storage medium, the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement steps of the electric vehicle aggregation and bidding method for peak-shaving auxiliary services as described in the embodiment 1.

[0288] Those skilled in the art will understand that the embodiments of the present disclosure can be provided as a method, a system or a computer program product. Therefore, the present disclosure can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Moreover, the present disclosure can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0289] The present disclosure is described with reference to flowcharts and / or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present disclosure. It should be understood that each flow and / or block in the flowcharts and / or block diagrams and the combination of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device that implements the flow Figure 1 one or more flows and / or blocks Figure 1 an apparatus that performs the functions specified in the flow or flows and / or block or blocks.

[0290] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce a manufactured product including instruction apparatus, which implements the flow Figure 1 one or more flows and / or blocks Figure 1 an apparatus that performs the functions specified in the flow or flows and / or block or blocks.

[0291] These computer program instructions can also be loaded into a computer or other programmable data processing device, so that a series of operation steps are performed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide a process for implementing the flow Figure 1 one or more flows and / or blocks Figure 1 an apparatus that performs the functions specified in the flow or flows and / or block or blocks.

Claims

1. A method for electric vehicle aggregation and bidding for peak-shaving ancillary services, characterized in that, The method comprises the following steps: A single electric vehicle dynamic model is established based on electric vehicle operation parameter data; An electric vehicle cluster aggregation dynamic model is obtained by solving the single electric vehicle dynamic model through an aggregation algorithm, and the model is specifically as follows: where E i (t) is the state of charge of the ith electric vehicle at time t, P i ch (t) is the state of charge of the ith electric vehicle at time t, P i dis (t) is the state of charge of the ith electric vehicle at time t, P (t) is the state of charge of the ith electric vehicle at time t, P E i (t) is the state of charge of the ith electric vehicle at time t, P (t) is the state of charge of the ith electric vehicle at time t, P i (t) is the state of charge of the ith electric vehicle at time t, P i,a (t) is the state of charge of the ith electric vehicle at time t, P i,l (t) is the state of charge of the ith electric vehicle at time t, P (t) is the state of charge of the ith electric vehicle at time t, P A double-layer strategy model of an electric vehicle aggregator participating in a peak regulation market is established based on the electric vehicle cluster aggregation dynamic model; The double-layer strategy model of the electric vehicle aggregator participating in the peak regulation market is converted into a single-layer mathematical optimization function with an equality constraint through Karush-Kuhn-Tucker conditions and a strong duality principle, and the function is solved to obtain a power purchase plan of the electric vehicle aggregator participating in the power market and a quantity and a price of the electric vehicle aggregator participating in the peak regulation auxiliary service market.

2. The method of claim 1, wherein, The single electric vehicle dynamic model is converted into a polyhedral half-space form, and the model is specifically as follows: The electric vehicle cluster aggregation dynamic model is obtained by solving the single electric vehicle dynamic model through an aggregation algorithm, and the model is specifically as follows: where P i denotes the D time node power of the i-th electric vehicle, is the set of power profiles of the i-th electric vehicle, denoting the range of operable power of the i-th electric vehicle; Half-space representation of a single electric vehicle dynamic model i P i <= b i , A i , b i As follows: Wherein, E i (0) is the electric energy state of the i-th electric vehicle at time 0, D≥t i,l -t i,a , D is the number of times of aggregation by the aggregator, and the specific matrix form is as follows: The electric vehicle cluster aggregation dynamic model is expressed in the form of a Minkowski sum of N electric vehicle dynamic models, and the model is specifically as follows: The NP-hard problem of N half-spaces is calculated by using an approximate solving method, and the problem is specifically as follows: wherein, is a cluster comprising N electric vehicles, is an aggregated dynamic model of the cluster of electric vehicles; The objective function is converted into a convex function in the form of a two-norm by using a triangle inequality of a Hausdorff distance and Lipschitz continuity, and the function is specifically as follows: Power feasible region of n electric vehicles approximated in the form with the same matrix Aggregated power feasible region of a cluster of electric vehicles may be approximated in the form The Hausdorff distance is used to measure the approximation quality The external and internal approximation problems can be described as follows, respectively: or wherein, Pout is an outer approximation of the power feasible region for EV i, Pin is an inner approximation of the power feasible region for EV i; According to Farkas' lemma, a set inclusion constraint in the approximate optimization problem is converted into a linear constraint and a nonlinear constraint; for the nonlinear constraint, a basic polyhedral scaling algorithm is used for solving, and the constraint is specifically as follows: right Simplify: Constructing a non-empty set: According to the Hausdorff distance triangle inequality, we have: Because L(A0) is the Lipschitz constant of the matrix A0, and thus the objective function can be equivalently converted into a convex function in the form of two-norm The constraint condition of the outer approximation is a linear constraint, and the constraint is specifically as follows: The constraint condition of the inner approximation is a nonlinear constraint, and the constraint is specifically as follows. The double-layer strategy model of the electric vehicle aggregator participating in the peak regulation market is established based on the electric vehicle cluster aggregation dynamic model, and the model is specifically as follows: For the case that the inner approximation condition is a nonlinear constraint, the outer approximation polytope is solved The basic polytope is solved by the basic polytope scaling algorithm, and the inner approximation problem is simplified as where s i = 1 / φ i , r i = -s i , ψ i ; Adopting Approximating an accurate dynamic model of an electric vehicle The aggregated dynamic model of the cluster of electric vehicles is then:

3. The method of claim 2, wherein, An electric vehicle aggregator participating in a power market model is established based on the electric vehicle cluster aggregation dynamic model, and the model is specifically as follows: An electric vehicle aggregator participating in a peak regulation auxiliary service market model is established based on the electric vehicle aggregator participating in the power market model, and the model is specifically as follows: wherein, is the decision variable, is the charging power of the electric vehicle aggregator, λ EM (t) is the time-of-use price published by the electricity market; An electric vehicle aggregator participating in a peak regulation auxiliary service market model is established based on the electric vehicle aggregator participating in the power market model, and the model is specifically as follows: λ EM (t)P agg (t) is the purchase cost of the power market for the aggregator, is the optimal power curve of the electric vehicle aggregator calculated according to the calculation, is the power adjustment amount of the electric vehicle aggregator, is the income of the electric vehicle aggregator participating in the peak shaving auxiliary service market, the peak shaving auxiliary service market settles the income according to the adjustment amount of the peak shaving participant, and λ(t) is the clearing price of the peak shaving auxiliary service market. The constraint condition is as follows: Objective function: The peak regulation supply-demand balance constraint corresponds to a dual variable λ(t) which is a peak regulation auxiliary service market clearing price, and the constraint is as follows: the maximum power reachable for each time node; the minimum power reachable for each time node, the up and down range of the optimal power curve difference for only participating in the electricity market, whose corresponding dual variable is as follows: The adjustable range of the large-scale peak-shaving device is constrained, and the corresponding dual variable is as follows: ​ 4. The method of claim 3, wherein, The double-layer strategy model of the electric vehicle aggregator participating in the peak shaving market is converted into a single-layer mathematical optimization function with equality constraints by using the Karush-Kuhn-Tucker condition and the strong duality principle, and the purchase plan of the electric vehicle aggregator participating in the power market and the quantity and price of the electric vehicle aggregator participating in the peak shaving auxiliary service market are obtained, which specifically includes: The Lagrange function is constructed for the lower-layer strategy model of the electric vehicle aggregator participating in the peak shaving market, and the objective function and the constraint condition of the lower-layer model are combined with the Lagrange multiplier, as shown in the following formula: The KKT condition is applied to the lower-layer strategy model of the electric vehicle aggregator participating in the peak shaving market, as shown in the following formula: The complementary slackness condition in the KKT condition is as follows: According to the KKT condition and the strong duality principle, the double-layer strategy model of the electric vehicle aggregator participating in the peak shaving auxiliary service market is converted into a single-layer mathematical optimization problem with equality constraints, as shown in the following formula: The objective function can be obtained according to the complementary slackness condition: According to the KKT condition, the following formula can be obtained: The KKT condition result is substituted into the complementary slackness condition result, and the following formula can be obtained: According to the strong duality principle, the following formula can be obtained: Therefore, the objective function of the single-layer model of the electric vehicle aggregator participating in the peak shaving market is finally obtained as follows: The constraint condition includes the adjustable range constraint of the electric vehicle aggregator, the reported peak shaving range constraint, the adjustable range constraint of the peak shaving device, and the KKT condition constraint, as shown in the following formula: The complementary slackness condition is linearized as follows: A0P agg (t) <= b agg The nonlinear function sign(P o (t)) is linearized with the complementary relaxed condition containing a bilinear term, respectively, as follows: The sign function sign(P o (t)) is linearized:

5. A system for implementing the peak shaving auxiliary service-oriented electric vehicle aggregation and bidding method according to claim 1, comprising: By solving the bi-level strategic model of the aggregator participating in the peak shaving market into a mixed integer linear programming problem, where the continuous decision variables are P agg (t), P o (t), α agg (t), λ(t), The decision variables are Obtain the purchase plan of the power market and the bid of the peak shaving auxiliary service market and the bid a dynamic model construction module of an electric vehicle, configured to establish a dynamic model of a single electric vehicle based on electric vehicle operation parameter data; a dynamic model construction module of an electric vehicle cluster aggregation, configured to establish a dynamic model of an electric vehicle cluster aggregation by using an aggregation algorithm to approximately optimize and solve based on the dynamic model of a single electric vehicle; a double-layer strategy model construction module of an aggregator participating in a peak shaving market, configured to establish a double-layer strategy model of an electric vehicle aggregator participating in a peak shaving market based on the dynamic model of an electric vehicle cluster aggregation; a double-layer strategy model solving module of an aggregator participating in a peak shaving market, configured to convert the double-layer strategy model of the electric vehicle aggregator participating in the peak shaving market into a single-layer mathematical optimization function with equality constraints by using the Karush-Kuhn-Tucker condition and the strong duality principle, and obtain the purchase plan of the electric vehicle aggregator participating in the power market and the quantity and price of the electric vehicle aggregator participating in the peak shaving auxiliary service market. The dynamic model of a single electric vehicle is established based on electric vehicle operation parameter data, which includes the power, state of charge, and charging and discharging time limit of the electric vehicle, and specifically includes the following:

6. The electric vehicle aggregation and bidding system for peak shifting ancillary services according to claim 5, wherein, The dynamic model of a single electric vehicle is established based on electric vehicle operation parameter data, which includes the power, state of charge, and charging and discharging time limit of the electric vehicle, and specifically includes the following: The dynamic model of a single electric vehicle is established based on electric vehicle operation parameter data, which includes the power, state of charge, and charging and discharging time limit of the electric vehicle, and specifically includes the following: where E i (t) is the state of charge of the ith electric vehicle at time t, P i ch (t) is the state of charge of the ith electric vehicle at time t, P i dis (t) is the state of charge of the ith electric vehicle at time t, P (t) is the state of charge of the ith electric vehicle at time t, P E i (t) is the state of charge of the ith electric vehicle at time t, P (t) is the state of charge of the ith electric vehicle at time t, P i (t) is the state of charge of the ith electric vehicle at time t, P i,a (t) is the state of charge of the ith electric vehicle at time t, P i,l (t) is the state of charge of the ith electric vehicle at time t, P (t) is the state of charge of the ith electric vehicle at time t, P ​ The dynamic model of a single electric vehicle is converted into a polyhedral half-space form, specifically as follows: where P i represents the D time node power of the i-th electric vehicle, is the set of power curves of the i-th electric vehicle, representing the range of power that the i-th electric vehicle can operate; Half-space representation of a single electric vehicle dynamic model i P i <= b i , A i , b i As follows: Wherein, E i (0) is the electric energy state of the i-th electric vehicle at time 0, D≥t i,l -t i,a , D is the number of times of aggregation by the aggregator, and the specific matrix form is as follows: Based on the dynamic model of a single electric vehicle and its polyhedral half-space form, the aggregated dynamic model of a cluster of electric vehicles is solved approximately by using the Minkowski sum and M-Sum, specifically as follows: The aggregated dynamic model of a cluster of electric vehicles is expressed in the form of the Minkowski sum and M-Sum of N dynamic models of electric vehicles as follows: wherein, is a cluster comprising N electric vehicles, is an aggregated dynamic model of the cluster of electric vehicles; The NP-hard problem of calculating N half-spaces is solved approximately, specifically as follows: Power feasible region of n electric vehicles approximated in the form with the same matrix Aggregated power feasible region of a cluster of electric vehicles may be approximated in the form ​ The Hausdorff distance is used to measure the approximation quality The external and internal approximation problems can be described as follows, respectively: or wherein, Pout is an outer approximation of the power feasible region for EV i, Pin is an inner approximation of the power feasible region for EV i; By using the triangle inequality of Hausdorff distance and Lipschitz continuity, the objective function is converted into a convex function in the form of two norms, specifically as follows: right Simplify: Constructing a non-empty set: According to the Hausdorff distance triangle inequality, we have: Because L(A0) is the Lipschitz constant of the matrix A0, and thus the objective function can be equivalently converted into a convex function in the form of two-norm According to Farkas' lemma, the set inclusion constraint in the approximate optimization problem is converted into a linear constraint and a nonlinear constraint; for the nonlinear constraint, a basic polyhedral scaling algorithm is used for solving, specifically as follows: The constraint condition of the outer approximation is a linear constraint, as follows: The constraint condition of the inner approximation is a nonlinear constraint, as follows: For the case that the inner approximation condition is a nonlinear constraint, the outer approximation polytope is solved The basic polytope is solved by the basic polytope scaling algorithm, and the inner approximation problem is simplified as where s i = 1 / φ i , r i = -s i , ψ i ; Adopting Approximating an accurate dynamic model of an electric vehicle The aggregated dynamic model of the cluster of electric vehicles is then:

7. The electric vehicle aggregation and bidding system for peak shifting ancillary services according to claim 6, wherein, Based on the aggregated dynamic model of a cluster of electric vehicles, a bi-level strategy model of an electric vehicle aggregator participating in the peak regulation market is established, including: Based on the aggregated dynamic model of a cluster of electric vehicles, an electric vehicle aggregator participating in the electricity market model is established, in which the charging cost of the electric vehicle aggregator is minimized within the feasible region of the electric vehicle cluster, and the expression is as follows: wherein, is the decision variable, is the charging power of the electric vehicle aggregator, λ EM (t) is the time-of-use price published by the electricity market; Based on the electric vehicle aggregator participating in the electricity market model, an upper-level strategy model of the electric vehicle aggregator participating in the peak regulation market is established, in which the quantity and price of the electric vehicle aggregator participating in the peak regulation auxiliary service market are reported while the overall cost of the electric vehicle aggregator is minimized, and the expression is as follows: λ EM (t)P agg (t) is the purchase cost of the power market for the aggregator, is the optimal power curve of the electric vehicle aggregator calculated according to the calculation, is the power adjustment amount of the electric vehicle aggregator, is the income of the electric vehicle aggregator participating in the peak shaving auxiliary service market, the peak shaving auxiliary service market settles the income according to the adjustment amount of the peak shaving participant, and λ(t) is the clearing price of the peak shaving auxiliary service market. An electric vehicle aggregator participating in the peak regulation market is established, in which the electric vehicle aggregator and a large battery participate in the peak regulation auxiliary service market, and the overall peak regulation cost is minimized under the premise of meeting the peak regulation demand, and the expression is as follows: Objective function: Constraint conditions: the maximum power reachable for each time node; the minimum power reachable for each time node, the up and down range of the optimal power curve difference for only participating in the electricity market, whose corresponding dual variable is as follows: The adjustable range of the large-scale peak-shaving device is constrained, and the corresponding dual variable is as follows: The peak regulation supply and demand balance constraint has a corresponding dual variable λ(t), which is the clearing price of the peak regulation auxiliary service market, as follows:

8. The electric vehicle aggregation and bidding system for peak-shaving ancillary services according to claim 7, wherein, The bi-level strategy model of the electric vehicle aggregator participating in the peak regulation market is converted into a single-level mathematical optimization function with equality constraints by using the Karush-Kuhn-Tucker condition and the strong duality principle, and the purchase plan of the electric vehicle aggregator participating in the electricity market and the quantity and price of the electric vehicle aggregator participating in the peak regulation auxiliary service market are obtained, specifically including: A Lagrangian function is constructed for the lower-level strategy model of the electric vehicle aggregator participating in the peak regulation market, and the objective function and constraint conditions of the lower-level model are combined with the Lagrange multiplier, as follows: The KKT condition is applied to the lower-level strategy model of the electric vehicle aggregator participating in the peak regulation market, as follows: The complementary slackness condition in the KKT condition is as follows: By using the strong duality principle, the optimal solution of the lower-level strategy model of the electric vehicle aggregator participating in the peak regulation market is converted into the constraint condition of the upper-level strategy model of the electric vehicle aggregator participating in the peak regulation market, and the bi-level model is converted into a single-level model, as follows: According to KKT condition and strong duality principle, the double-layer strategy model of the electric vehicle aggregator participating in the peak shaving auxiliary service market is converted into a single-layer mathematical optimization problem with equality constraints, as follows: The objective function can be obtained according to the complementary relaxation condition: According to the KKT condition, the following can be obtained: The KKT condition result is substituted into the complementary relaxation condition result to obtain the following: The strong duality principle is substituted to obtain the following: Therefore, the objective function of the single-layer model of the electric vehicle aggregator participating in the peak shaving market is finally obtained as follows: The constraint conditions include the adjustable range constraint of the electric vehicle aggregator, the reported peak shaving range constraint, the adjustable range constraint of the peak shaving device, and the KKT condition constraint, as follows: A0P agg (t) <= b agg The nonlinear function sign(P o (t)) is linearized with the complementary relaxed condition with a bilinear term, respectively, as follows: The linearization of the sign function sign(P o (t)) is given by The complementary relaxation condition is linearized: By solving the bi-level strategic model of the aggregator participating in the peak shaving market into a mixed integer linear programming problem, where the continuous decision variables are P agg (t), P o (t), α agg (t), λ(t), The decision variables are Obtain the purchase plan of the power market and the bid of the peak shaving auxiliary service market and the price 9. A computer storage medium storing a computer program, the computer program comprising instructions, which, when executed by a computer, cause the computer to perform the method according to any one of claims 1 to 8. The computer program is executed by the processor to realize the steps of the electric vehicle aggregation and bidding method for peak shaving auxiliary services according to any one of claims 1-4.

Images