Coordinated bidding method for virtual power plants to participate in electricity market and carbon trading market

Abstract

The present invention relates to a collaborative bidding method for virtual power plants to participate in the electricity market and the carbon trading market, which is characterized by: comprising four steps, the operation mechanism and structure analysis of virtual power plants; the bidding strategy of virtual power plants in the day-ahead market and the regulation market; the virtual power plant bidding model and the robust optimization bidding model of virtual power plants. Beneficial effects: The present invention can provide technical support for virtual power plants to participate in the collaborative bidding in the electricity market and the carbon trading market, solves the problems of unclear cost-benefit structure and risk preference that virtual power plants need to face in the collaborative bidding, meets the risk-benefit trade-off principle, is suitable for scenarios where virtual power plants with multiple risk preferences participate in the collaborative bidding in the electricity market and the carbon trading market, and has wide practicality.

Description

Technical Field

[0001] The present invention belongs to the technical field of electricity market, and in particular relates to a collaborative bidding method for a virtual power plant to participate in an electricity market and a carbon trading market. Background Art

[0002] The diversity of loads and the high penetration of renewable energy access have brought new challenges to the economic operation, safety and stability of traditional power systems. As a new type of coordinated control and energy management technology for distributed power sources, virtual power plants (VPPs) can aggregate various distributed power sources in a region into a whole based on Internet communication technology, and participate in the bidding in the power market and carbon trading market in a unified manner. Multi-regional virtual power plants can also achieve inter-regional energy flow complementarity, reduce the difficulty of grid dispatching, improve the utilization efficiency and absorption level of renewable energy, and promote the transformation of the grid from a single energy consumption structure to a diversified energy consumption structure.

[0003] Various distributed power sources, controllable loads and energy storage devices can be aggregated in a virtual power plant. Combined with various market price information and the output characteristics of various units, virtual power plants play multiple roles in the power market. Not only can it participate in market bidding as a power supplier to sell electricity, but it can also purchase electricity from the market as a load to meet its own power needs. With a diversified energy structure, flexible energy supply forms and various energy storage devices, virtual power plants participate in regulating market bidding by dispatching distributed power sources, controllable loads and energy storage devices, and provide backup capacity for the power system.

[0004] At present, there are many studies on the bidding strategies for wind turbines to participate in the electricity market, and a collaborative bidding method considering wind turbines and fixed energy storage devices has been proposed. Electric vehicles have commuting and energy storage functions, which makes electric vehicles as an energy storage device have certain cost advantages. Using the energy storage function of electric vehicles to provide backup capacity for the power system is a promising commercial operation service model, namely the vehicle-to-grid (V2G) model. However, there are few studies considering the joint bidding of electric vehicles and wind turbines.

[0005] Due to the intermittent and random nature of renewable energy generation, virtual power plants with high penetration of renewable energy will face high trading risks in the power market. Domestic and foreign studies have shown that using demand response resources as alternative resources on the power supply side is an important means to alleviate the trading risks of virtual power plants in the power market. In recent years, the demand response market has also been considered an important part of the future power market and has attracted widespread attention. Given that demand response transactions can stabilize the uncertainty of renewable power output, reduce trading risks, and improve economic benefits. Domestic and foreign studies have used demand response as an alternative resource on the power supply side of virtual power plants and established a two-layer random scheduling optimization model for virtual power plants. The patent document with application publication number CN 112465248A discloses a method and system for optimizing the operation and profit distribution of virtual power plants considering carbon trading. The method includes: establishing a virtual power plant optimization model with the goal of maximizing the sum of the virtual power plant's carbon market revenue and electricity market revenue; solving the virtual power plant optimization model to obtain the bidding volume in the day-ahead market; according to the bidding volume in the day-ahead market, using the Shapley value calculation method to determine the revenue distribution of virtual power plant members; wherein the virtual power plant members include gas turbine units, energy storage devices, wind turbine units and photovoltaic units. The method and system of the invention can provide theoretical guidance for the behavioral decision-making of multiple demand response resources participating in the electricity market and carbon market.

[0006] In addition, virtual power plants can also participate in the bidding in the carbon trading market. At present, domestic and foreign research has been conducted on the bidding strategies of virtual power plants participating in the carbon trading market, and a virtual power plant economic dispatch model considering the carbon trading mechanism has been proposed, aiming to maximize the income of virtual power plants.

[0007] From the above analysis, it can be seen that the output, load and market price of virtual power plants are uncertain, and the uncertainty in these three aspects is inevitable. When studying the bidding strategy of virtual power plants participating in the power market, it is necessary to consider the uncertainty of the number of dispatchable electric vehicles and the output of wind turbines. As a method for dealing with uncertain factors, robust optimization has attracted widespread attention in the fields of natural science, engineering technology, economic management, etc. in recent years, and its application fields are becoming more and more extensive. At present, the bidding strategies of virtual power plants are mostly based on the power trading analysis model of the power market, and carbon emission rights are not introduced into the power market for consideration. With the increase in the penetration rate of distributed power sources and the improvement of load controllability, it is of great significance to fully tap the controllable resources of virtual power plants and build a collaborative bidding strategy model for virtual power plants to participate in the power market and carbon trading market, which is of great significance to improve the economic efficiency of virtual power plants. In addition, the application research of demand response resources in virtual power plants mainly focuses on the economic operation and optimal scheduling of virtual power plants, and lacks research considering the impact of demand response market on the bidding strategy of virtual power plants. Summary of the invention

[0008] The purpose of the present invention is to overcome the shortcomings of the above-mentioned technology and to provide a collaborative bidding method for virtual power plants to participate in the electricity market and the carbon trading market. The method can simultaneously participate in the day-ahead electricity market, the intraday demand response market, the regulation market, the real-time electricity market and the carbon trading market, effectively improve the profits of virtual power plants, reduce the risks caused by the uncertainty of renewable energy generation, load and electricity prices, and meet the collaborative bidding strategies of virtual power plants with different risk preferences.

[0009] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: a collaborative bidding method for a virtual power plant to participate in the electricity market and the carbon trading market, characterized in that it comprises four steps:

[0010] 1. Analysis of the operation mechanism and structure of virtual power plants: Taking virtual power plants with operating structures consisting of power generation units, energy storage units and demand response units as the object, a robust optimization bidding strategy for virtual power plants participating in the day-ahead power market, intraday demand response market, regulation market, real-time power market and carbon trading market is proposed; by analyzing the business model of virtual power plants, the planning of participating in power market bidding, and the integration of power market and carbon trading market, the diversified income and cost structure of virtual power plants participating in the coordinated bidding of power market and carbon trading market is clarified;

[0011] 2. Bidding strategies of virtual power plants in the day-ahead market and regulation market: By establishing revenue and cost models for wind turbines, electric vehicles and gas turbines, clarify the bidding strategies of virtual power plants participating in the day-ahead power market and regulation market; establish a quantitative model for the uncertainty of virtual power plants, and reduce the bidding risks of virtual power plants caused by the uncertainty of renewable energy generation through the demand response market;

[0012] 3. Virtual power plant bidding model: Introduce carbon trading mechanism, aim at maximizing the profit of virtual power plants, and optimize the collaborative bidding strategy of virtual power plants participating in the electricity market and carbon trading market;

[0013] 4. Robust optimization bidding model for virtual power plants: Considering the different risk preferences of virtual power plant bidding, a robust optimization model for the collaborative bidding strategy of virtual power plants participating in the electricity market and carbon trading market is established, further improving the profits of virtual power plants with different risk preferences participating in the collaborative bidding in the electricity market and carbon trading market.

[0014] Furthermore, the specific steps of the virtual power plant operation mechanism and structure analysis are as follows:

[0015] (1) Business model of virtual power plant in the power distribution market

[0016] The distribution market business model of virtual power plants includes four stages:

[0017] ① Based on the forecast information of distributed renewable energy, the virtual power plant aggregates its internal resources according to the output range and operating cost to obtain the output curve and cost characteristics of the aggregated flexibility resources;

[0018] ② The virtual power plant submits bidding prices and quantities of multiple power commodities to the distribution network operator;

[0019] ③ The distribution network operator collects the bidding information of various operators, considers the bidding output and cost characteristics of various units, proposes market clearing strategies and determines the winning bid price and winning bid quantity of virtual power plants based on the demand for various power commodities and the constraints of system safe operation;

[0020] ④ The virtual power plant optimizes the internal resource scheduling plan according to the market clearing results and follows the scheduling instructions issued by the distribution network operator. The virtual power plant designs the scheduling plan of its internal flexibility resources according to the scheduling plan curve and various electricity commodity prices issued by the distribution network operator, and tracks the scheduling instructions issued by the distribution network operator.

[0021] (2) Bidding rules for virtual power plants to participate in the electricity market

[0022] The bidding rules for virtual power plants are as follows: based on the participation of virtual power plants with high penetration of renewable energy in the day-ahead market, intraday demand response market, regulation market and real-time market,

[0023] Virtual power plants need to predict their actual power generation 12 to 36 hours in advance and submit bidding information for the next 24 hours to distribution network operators before the day-ahead market closes, i.e., the power generation price-power generation capacity curve; set the bidding price of virtual power plants to 0, and optimize their bidding power only based on the predicted market settlement price; the regulation market is closed one hour before the real-time market opens, during which time virtual power plants submit bids for regulation reserves and spinning reserves to distribution network operators. After the power market transaction ends on the day, the actual power generation deviation of virtual power plants will be settled based on the real-time electricity price. The bidding power generation is used to study the bidding strategy of virtual power plants. The hourly power generation and hourly power generation of virtual power plants have the same data, i.e., hourly power generation = hourly power generation × 1 hour;

[0024] (3) Integration characteristics of the electricity market and carbon trading market in which virtual power plants participate

[0025] In the operation structure of the virtual power plant: the power generation unit is mainly composed of wind turbines and gas turbines, the energy storage unit is an electric vehicle, and the demand response unit is a local load and an external demand response provider. Each unit transmits daily operation data to the control center of the virtual power plant through the data layer. The control center formulates strategies and dispatches the output of the power generation unit equipment, the response capacity of the energy storage unit and the demand response unit according to the current profit and carbon emission targets. The power generation units, energy storage units and demand response units in the aggregation area of ​​the virtual power plant are the basis for building an integrated power market and carbon trading market. The process of each unit in the virtual power plant participating in market transactions is also the process of transferring electricity rights and carbon emission rights at the same time.

[0026] 2) Bidding strategies of virtual power plants in the day-ahead market and regulation market

[0027] (1) Revenue function of wind turbines

[0028] After winning the bid in the day-ahead market, when the wind turbine is dispatched to generate electricity in the real-time market, during the T period, the day-ahead market wind turbine bidding revenue R WT,da , as shown in equation (1)

[0029]

[0030] Where i is the node number of the distribution network; B is the set of distribution network nodes; λ da,e,t is the day-ahead electricity price; is the set of wind turbines connected to node i; P i,wt,t is the day-ahead bidding output of the wind turbine at time t; Δt represents the time interval;

[0031] (2) Electric vehicles provide backup capacity for distribution networks

[0032] First, the income function

[0033] When electric vehicles participate in the regulation market bidding, the bidding capacity of upward reserve, downward reserve and spinning reserve are expressed as and

[0034] When electric vehicles provide spare capacity, the dispatch ratio of spare capacity is statistically analyzed through historical dispatch data. The expected dispatch ratios of electric vehicles for upward reserve capacity, downward reserve capacity, and rotating reserve capacity are expressed as and As shown in equation (6-8)

[0035]

[0036]

[0037] In the formula, and are the actual dispatch outputs of the upward reserve capacity, downward reserve capacity and spinning reserve capacity of the electric vehicle connected to the node i at time t in the real-time market; and are the upward reserve capacity, downward reserve capacity and spinning reserve capacity of the ev-th electric vehicle connected to the node i at time t, respectively; E(·) represents the expected value;

[0038] The exchange power P between electric vehicles and distribution grid EV,t , as shown in equation (9)

[0039]

[0040]

[0041] In the formula, To meet the charging power required for electric vehicles; and are the total charging and discharging power of the electric vehicle respectively;

[0042] Virtual power plant according to electricity price λ ev,sub Charging fees are charged to dispatchable electric vehicles, and the total income of electric vehicles in the regulation market during period T is R EV,rg ,

[0043] As shown in equation (11),

[0044]

[0045] In the formula, To meet the charging power demand of the evth electric vehicle connected to node i at time t; RD,t ,λ RU,t ,λ RR,t They are the capacity prices of downward reserve, upward reserve and spinning reserve in the regulation market;

[0046] Second, the cost function

[0047] The cost of electric vehicles participating in the regulation market bidding consists of two parts, as shown in equation (12):

[0048] C EV =C EV,pur +C EV,los (12)

[0049] First, the cost of electricity purchased by the virtual power plant for charging electric vehicles is C EV,pur , as shown in equation (13)

[0050]

[0051] In the formula, λrt,e,t Real-time electricity price.

[0052] Secondly, the dispatchable electric vehicles of the virtual power plant use the same type of batteries, and the loss cost of the electric vehicle battery is shown in equation (14).

[0053]

[0054] Where P EV,bat is the rated capacity of the electric vehicle battery; C bat is the purchase cost coefficient of electric vehicle batteries, RMB / kWh. η dc is the discharge efficiency of the electric vehicle battery;

[0055] (3) Gas Generator

[0056] First, the income function

[0057] The gas-fired unit that wins the bid in the day-ahead market will be dispatched to generate electricity in the real-time market. During the T period, the bidding income of the gas-fired unit in the day-ahead market is R GT,da , as shown in equation (16)

[0058]

[0059] In the formula, is the set of gas units connected to node i; P i,gt,t is the bid output of the gas generating units at each time period t a day ago;

[0060] When gas units participate in the market bidding, the bidding output of upward reserve, downward reserve and spinning reserve is and express, and They are the actual dispatch value of the gas unit's upward reserve, the actual dispatch value of the downward reserve, and the actual dispatch value of the rotating reserve.

[0061] After the gas unit wins the bid in the regulation market, the total revenue R GT,rg , as shown in equation (17)

[0062]

[0063] Second, the cost function

[0064] The cost of purchasing natural gas from the gas market for the virtual power plant GT,pur , as shown in equation (18).

[0065]

[0066] In the formula, λ gasis the natural gas price; η i,gt,t is the power generation efficiency of the gas unit; LHV is the lower heating value of natural gas, kWh / m 3 ;

[0067] 3) Virtual power plant bidding model

[0068] (1) Uncertainty modeling

[0069] Uncertain factors include wind turbine output, local user load demand and market prices, where market prices include day-ahead market prices, adjustment market prices, intraday demand response market prices and real-time market prices.

[0070] ② Fan output

[0071] In the day-ahead market, the interval constraint method is used to set the upper and lower limits of the wind turbine output. The wind turbine output P at time t is WT,t The upper and lower limits of the uncertainty parameters P up,WT,t and P low,WT,t Indicates that at any time t, there is only one uncertain parameter in the upper and lower limits of the wind turbine bidding output, which are P up,WT,t and P low,WT,t ; The upper and lower limits of wind turbine output follow a normal distribution, as shown in equation (19-21)

[0072]

[0073]

[0074] σ Wind ={σ up,WT,t ,σ low,WT,t} Wind ={μ up,WT,t ,μ low,WT,t} (twenty one)

[0075] In the formula, μ up,WT,t , σ up,WT,t YesP up,WT,t The expected value and standard deviation of low,WT,t , σ low,WT,t YesP low,WT,t The expected value and standard deviation of

[0076] The opportunity constraint can be used to convert the wind turbine bidding output constraint into an inequality constraint, as shown in equations (22-23):

[0077] pr{P WT,t ≤P up,WT,t}≥ε up,WT (twenty two)

[0078] pr{P WT,t≥P low,WT,t}≥ε low,WT (twenty three)

[0079] In the formula, ε up,WT and ε low,WT the probability of satisfying the chance constraint;

[0080] ②Local user load demand

[0081] The virtual power plant needs to meet the load demand of local users and is predicted based on historical load data, as shown in equation (24):

[0082]

[0083] In the formula, It predicts local user load based on historical load data through the difference autoregressive moving average model; are the upper and lower limits of the prediction error, respectively, which follow the normal distribution;

[0084] ③ Market transaction price

[0085] According to historical data, the day-ahead electricity price, demand response market price, reserve capacity price and real-time electricity price are all within the range of Internal changes, among which represents the predicted value of price; Indicates the radius of the price fluctuation range;

[0086] (2) Objective function

[0087] The goal of the virtual power plant bidding strategy is to maximize the virtual power plant profit I, as shown in equation (25):

[0088] maxI=R rev -C WT,pl -C EV -C GT,pur -C DR (25)

[0089] R rev =R WT,da +R GT,da +R EV,rg +R GT,rg +R LU +R CT (26)

[0090] In the formula, R revis the income obtained by the virtual power plant, including: the income of wind turbines and gas turbines in the day-ahead market in equations (1) and (16), the income of electric vehicles and gas turbines in the regulation market in equations (11) and (17), the income from selling electricity to local users in equation (27), and the income from carbon emission rights in the carbon trading market in equation (28); C WT,pl is the economic penalty caused by the deviation between the actual output of the wind turbine in equation (2) and the bid output, C EV is the electric vehicle cost in equation (12), C GT,pur is the natural gas purchase cost of the gas-fired unit in equation (18), C DR is the purchase cost of load reduction in the intraday demand response market of equation (30);

[0091] (3) Constraints

[0092] First, wind turbine operation constraints

[0093] The bid output and actual output of a wind turbine should be between its minimum and maximum output power, as shown in equations (39-40)

[0094]

[0095]

[0096] In the formula, and are the minimum and maximum output power of the wind turbine respectively;

[0097] Second, electric vehicle operation constraints

[0098] The states of electric vehicles include charging, discharging and standby, and need to satisfy the constraints of equation (41)

[0099]

[0100] In the formula, u dc,t is the discharge state of the electric vehicle. c,t Is the charging status of the electric vehicle

[0101] The electric vehicle SOC at time t is shown in equation (42). The electric vehicle SOC needs to satisfy the upper and lower capacity constraints, as shown in equation (43):

[0102]

[0103]

[0104] Where: SOC i,ev,t is the SOC value of the evth electric vehicle connected to node i in period t, SOCi,ev,0 is the initial SOC of the electric vehicle, They are the upper and lower limits of the electric vehicle SOC, P i,ev,bat is the rated capacity of the battery of the evth electric vehicle connected to node i;

[0105] Electric vehicles also need to meet the maximum charge and discharge power PEV,max constraint, as shown in equations (44-45)

[0106]

[0107]

[0108] Third, gas unit operation constraints

[0109] The bidding output of the gas unit should meet the upper and lower limits of its output power and the maximum ramp rate constraints, as shown in equations (46-49):

[0110]

[0111]

[0112]

[0113]

[0114] In the formula, is the maximum output power of the gas unit; It is the bidding capacity of gas units to be adjusted down for standby; and They are the maximum upward / downward ramp rates of the gas generator set;

[0115] Fourth, line constraints

[0116]

[0117]

[0118] In the formula, is the injected power of node i at time t; P ij,max , P ij,min are the upper and lower limits of the power of line ij respectively; h ij is the power distribution variation coefficient of line ij;

[0119] 4) Robust optimization bidding model for virtual power plants

[0120] First, the virtual power plant bidding model is transformed into a linear programming problem, and then into a robust linear programming model; by introducing decision variables, the constraint equations (3, 4, 9) are modified to form new constraint equations (52-55)

[0121]

[0122]

[0123]

[0124]

[0125] Convert equation (42) into a linear function equation (56) to calculate the SOC of an electric vehicle,

[0126]

[0127]

[0128] According to the robust optimization principle, the linear programming model is transformed into a robust optimization bidding model for virtual power plants, as detailed below:

[0129] max I ROM (58)

[0130]

[0131]

[0132]

[0133]

[0134]

[0135]

[0136]

[0137]

[0138]

[0139]

[0140]

[0141]

[0142]

[0143]

[0144] -y≤P≤y (73)

[0145]

[0146]

[0147]

[0148]

[0149]

[0150]

[0151]

[0152] Where P is the decision variable vector; y is the auxiliary decision variable vector introduced by the dual transformation; z ε is the upper quantile of the standard normal distribution; μ a,W,t and σ a,W,t They are The expected value and standard deviation of It follows a normal distribution.

[0153] In the robust optimization bidding model of virtual power plant bidding, the charging and discharging power constraints of electric vehicles satisfy equations (54) and (55), and the SOC constraints of electric vehicles satisfy equation (56); equations (77) and (78) are the upper and lower limit constraints of the bidding output of wind turbines, and equations (79) and (80) are the positive and negative deviation constraints of the bidding output of wind turbines; other constraints are shown in equations (43-51).

[0154] Furthermore, the cost function of wind turbines in the distribution market business model of the virtual power plant is

[0155] When there is a deviation between the bid output and the actual output of the wind turbine, it will be subject to economic penalties, as shown in equation (2). When the wind turbines are used to charge electric vehicles first, the remaining electricity is sold in the real-time market. By controlling the discharge of electric vehicles and purchasing load reduction from the demand response trading market, the deviation between the bid output and the actual output of wind turbines can be reduced, and the output fluctuation of wind turbines can be stabilized.

[0156]

[0157]

[0158]

[0159]

[0160] In the formula, and The actual output is greater than / less than the bid output of the wind turbine; u ec,t 、u bl,t They are the state variables of positive deviation / negative deviation of fan output respectively; and They are the charging and discharging power of electric vehicles to reduce the output deviation of wind turbines; wt,ec is the penalty coefficient for positive deviation of fan output, ω wt,ec <1;ω wt,bl is the penalty coefficient for negative deviation of fan output, ω wt,bl >1 is the actual output of the fan; is the set of electric vehicles connected to node i; γ DR is the set of demand response suppliers; P dr,t is the amount of load reduction purchased from the drth demand response supplier.

[0161] Furthermore, in the distribution market business model of the virtual power plant, when electric vehicles cannot be dispatched as planned, it is necessary to increase the charging and discharging power of other electric vehicles to compensate for the capacity loss. For a certain electric vehicle, when other electric vehicles cannot be dispatched as planned at time t, its charging and discharging power adjustment coefficient is shown in equation (15).

[0162]

[0163] In the formula, To adjust the coefficient, when the total number of electric vehicles reaches 10,000 or more, it can be regarded as a constant; is the probability that the evth electric vehicle connected to node i at time t cannot be dispatched as planned.

[0164] Furthermore, the objective function of the virtual power plant bidding model specifically includes income

[0165] ① Income from selling electricity to local users

[0166] The relationship between local user electricity consumption and electricity price: During the T period, the income from selling electricity to local users is the product of electricity consumption and electricity price, as shown in equation (27):

[0167]

[0168] In the formula, u lu,t is the ratio of local user electricity price to real-time electricity price, 0≤u i,lu,t ≤1; is the set of local users accessing node i; is the power consumption of the luth user at time t;

[0169] ② Income from the carbon trading market

[0170] The income of the virtual power plant from selling carbon emission rights in the carbon trading market is shown in equation (28):

[0171]

[0172]

[0173] In the formula, λ c,t is the price coefficient of carbon emission rights at time t, yuan / MW; P i,c,t is the bidding output of the wind turbine and gas turbine connected to node i at time t in the carbon trading market;

[0174] ③ Load reduction purchase cost in the demand response market

[0175] In the demand response market, virtual power plants purchase load reduction through bilateral negotiation and centralized bidding from demand response suppliers. The purchase cost is shown in equation (30):

[0176]

[0177] In the formula, γ DR is the set of demand response suppliers; C DR,cen is the cost of purchasing load reduction through centralized bidding transactions; C DR,bi is the cost of purchasing load reduction through bilaterally negotiated transactions;

[0178] The purchase cost of load reduction in centralized bidding is shown in equation (31):

[0179]

[0180] λ cen,dr,t =θ cen,dr λ rt,e,t (32)

[0181]

[0182]

[0183] In the formula, γ DR is the set of demand response providers; cen,dr,t is the bid price of the drth demand response supplier at time t; P cen,dr,t is the demand response bidding amount corresponding to the demand response bidding price; θ cen,dr is the adjustment coefficient, which can be adjusted by the distribution network operator; Q DR,cen,t is the total demand response provided through the centralized bidding model at time t; is the maximum bid amount of the drth demand response supplier;

[0184] The cost of purchasing load reduction through bilateral negotiation is given by equation (35):

[0185]

[0186]

[0187] Where λbi,dr,t is the bilaterally negotiated price of demand response, which can be agreed upon and determined in advance by both the supply and demand parties; Pbi,dr,t is the demand response purchase volume corresponding to the bilaterally negotiated price of demand response; QDR,Bi,t is the total demand response volume provided through the bilateral negotiation mode at time t.

[0188] The total load reduction purchased by the virtual power plant should meet the demand response capacity DRcap constraint that the demand response provider can provide, as shown in equation (37):

[0189]

[0190]

[0191] In the formula, DR cap The demand response capabilities that demand response providers can provide; P dr,t is the amount of load reduction purchased from the drth demand response supplier.

[0192] Furthermore, in the robust optimization bidding model of the virtual power plant, when the electric vehicle is charging, equation (56) is the same as equation (42). When the electric vehicle is discharging, the electric vehicle discharge amount calculated by equation (56) is less than the actual discharge amount, and Δerr is used. i,ev,t Indicates the deviation of discharge capacity. When the charge and discharge efficiency of electric vehicles is η c =η dc =0.95, Δerr i,ev,t Only accounts for the actual discharge 9.75% when η c =η dc =1.0, Δerr i,ev,t =0, replace equation (42) with equation (56).

[0193] Furthermore, in the robust optimization bidding model of the virtual power plant, in order to make the constraint violation probability not exceed κ, the robust control coefficient Γ1 needs to satisfy the constraint equation (81)

[0194]

[0195] In the formula, Φ -1is the inverse function of the cumulative distribution function of the standard normal distribution; n np is the number of uncertain parameters contained in constraint equation (59).

[0196] Furthermore, the collaborative bidding strategy of the virtual power plant participating in the electricity market and the carbon trading market will be constructed and solved by AMPL / CPLEX based on the MATLAB R2018b platform in an environment where the CPU is Inter(R)Core(TM) i7-8250U and the main frequency is 1.8GHz.

[0197] Beneficial effects: The present invention can provide technical support for virtual power plants to participate in the coordinated bidding of the electricity market and the carbon trading market, solve the problems of unclear cost-benefit structure and risk preference faced by the coordinated bidding of virtual power plants, meet the risk-benefit trade-off principle, and is suitable for scenarios where virtual power plants with multiple risk preferences participate in the coordinated bidding of the electricity market and the carbon trading market, and has wide practicality; it can fully utilize the business model of virtual power plants, the planning of participating in the bidding of the electricity market, and the integration of the electricity market and the carbon trading market, propose the diversified income and cost structure of virtual power plants participating in the coordinated bidding of the electricity market and the carbon trading market, and point out the direction for the bidding strategy of virtual power plants; it can alleviate the bidding risks caused by renewable energy generation through the electric vehicle energy storage and demand response market, and clearly Determine the bidding strategy for virtual power plants to participate in the day-ahead electricity market, intraday demand response market and regulation market; introduce the carbon trading mechanism, consider the different risk preferences of virtual power plant bidding, and build a robust optimization model for the collaborative bidding strategy of virtual power plants participating in the electricity market and the carbon trading market with the goal of maximizing the profits of virtual power plants. The final collaborative bidding strategy for virtual power plants to participate in the electricity market and the carbon trading market has high universality, good economic and social benefits, strong robustness, and can better balance risks and profits, reduce the investment risks brought about by the uncertainty of renewable energy generation, load and electricity prices, promote the consumption of renewable energy, reduce carbon dioxide emissions from power generation, promote the low-carbon transformation of the energy structure, and help achieve carbon peak as soon as possible, which has greater academic significance and engineering value. BRIEF DESCRIPTION OF THE DRAWINGS

[0198] Figure 1 The logical framework for virtual power plants to participate in the coordinated bidding in the electricity market and carbon trading market;

[0199] Figure 2a Business model framework for the power distribution market including virtual power plants;

[0200] Figure 2b It is the internal operation mode of the virtual power plant;

[0201] Figure 3 The distribution market structure and trading time frame including virtual power plants;

[0202] Figure 4 It is the integrated operation feature of the electricity market and carbon trading market in which virtual power plants participate;

[0203] Figure 5 It is the wind power output constraint and the expected actual output value;

[0204] Figure 6a is the electricity market price forecast curve;

[0205] Figure 6b is the expected value curve of the power market dispatch ratio;

[0206] Figure 7a is the local user load forecast curve;

[0207] Figure 7b It is the step curve of electricity price-electricity power for local users;

[0208] Figure 8a is the bilaterally negotiated price map for demand response suppliers;

[0209] Figure 8b is the bid price-bid quantity diagram of the demand response supplier in the centralized bidding model;

[0210] Fig. 9 It is the wind turbine bidding output change curve;

[0211] Fig.10a It is the electric vehicle charging and discharging power curve that smoothes the fluctuation of wind turbine power output;

[0212] Fig.10b It is the spinning reserve bidding capacity curve of electric vehicles in the regulation market;

[0213] Fig.10c It is the upward adjustment reserve bidding capacity curve of electric vehicles in the regulation market;

[0214] Fig.10d It is the downward adjustment reserve bidding capacity curve of electric vehicles in the regulation market;

[0215] Fig.11a is Wind =0.1μ Wind The curve diagram showing the impact of wind turbine output standard deviation and demand response capacity DRcap on wind turbine day-ahead market bidding power;

[0216] Fig.11b is Wind =0.2μ Wind The curve diagram showing the impact of wind turbine output standard deviation and demand response capacity DRcap on wind turbine day-ahead market bidding power;

[0217] Fig.12a It is a curve diagram showing the impact of wind turbine output standard deviation and demand response capacity DRcap on wind turbine bidding power generation;

[0218] Figure 12b This is the curve diagram showing the impact of wind turbine output standard deviation and demand response capacity DRcap on the profit of virtual power plants;

[0219] Fig.13 This is a curve diagram showing the impact of DRcap on local users’ electricity income under different wind turbine output standard deviation conditions;

[0220] Fig.14a The bidding output of gas generators when there is no carbon trading constraint (the bidding output power in the day-ahead market);

[0221] Fig.14b This is the bidding output situation of gas units when there is no carbon trading constraint (the bidding output power of the regulation market);

[0222] Fig.15a The bidding output of gas-fired units when there are carbon trading constraints (when bidding output power in the day-ahead market);

[0223] Fig.15b The bidding output of gas units when there are carbon trading constraints (when adjusting the market's bidding output power) DETAILED DESCRIPTION

[0224] In order to be able to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that, in the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other. In the following description, many specific details are set forth in order to fully understand the present invention, and the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in the art without making creative work are within the scope of protection of the present invention. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0225] The embodiments and drawings provide a detailed explanation of the collaborative bidding strategy of the virtual power plant of the present invention for participating in the electricity market and the carbon trading market.

[0226] At present, the output, load and market price of virtual power plants are uncertain, and the uncertainty in these three aspects is inevitable. When studying the bidding strategy of virtual power plants participating in the electricity market, it is necessary to consider the uncertainty of the number of dispatchable electric vehicles and the output of wind turbines. Moreover, the bidding strategies of virtual power plants are mostly based on the power trading analysis model of the power market, and carbon emission rights are not introduced into the power market for consideration. With the increase in the penetration rate of distributed power sources and the improvement of load controllability, it is of great significance to fully tap the controllable resources of virtual power plants and build a collaborative bidding strategy model for virtual power plants to participate in the electricity market and carbon trading market, which is of great significance to improving the economic efficiency of virtual power plants. In addition, the application research of demand response resources in virtual power plants mainly focuses on the economic operation and optimal scheduling of virtual power plants, and lacks research considering the impact of the demand response market on the bidding strategy of virtual power plants. To this end, the present invention takes a virtual power plant containing wind turbines, electric vehicles, gas turbines and demand response loads as the research object, analyzes the integrated characteristics of the power market and the carbon trading market, further considers the uncertainty of renewable energy generation, load and market price, and constructs a robust optimization model of the bidding strategy when the virtual power plant participates in the day-ahead market, the intraday demand response trading market, the regulation market, the real-time market and the carbon trading market. This makes the collaborative bidding strategy of the virtual power plant participating in the power market and the carbon trading market more practical, reduces the output fluctuation of renewable energy power generation, promotes the utilization of renewable energy, significantly reduces the bidding output of high-carbon emission generators, and improves the economic and social benefits of the virtual power plant.

[0227] The collaborative bidding strategy of the virtual power plant in the present invention for participating in the electricity market and the carbon trading market is divided into four steps, as shown in the attached figure. Figure 1 It includes four steps: 1) analysis of the operation mechanism and structure of virtual power plants, 2) bidding strategies of virtual power plants in the day-ahead market and regulation market, 3) bidding model of virtual power plants, and 4) robust optimization bidding model of virtual power plants, which are described in detail below.

[0228] 1) Virtual power plant operation mechanism and structure analysis

[0229] (1) Considering the business model of the distribution market for virtual power plants

[0230] The business model structure of the distribution market considering virtual power plants is shown in Figure 2. The business model includes three types of entities and two interaction layers. The three types of entities include distribution network operators, virtual power plants, and flexibility resource equipment managers. The three types of entities can be divided into two interaction layers. The first interaction layer is the information interaction and management control layer between virtual power plants and flexibility resource equipment managers. The second interaction layer is the market bidding and transaction layer between distribution network operators and virtual power plants.

[0231] The distribution market business model considering virtual power plants can be divided into the following four stages. First, based on the forecast information of distributed renewable energy, the virtual power plant aggregates its internal resources according to the output range and operating costs to obtain the output curve and cost characteristics of the aggregated flexibility resources. Second, the virtual power plant submits the bidding prices and quantities of various power commodities to the distribution network operator. Third, the distribution network operator collects the bidding information of various operators, considers the bidding output and cost characteristics of various units, and proposes a market clearing strategy and determines the winning price and winning quantity of the virtual power plant based on the various power commodity demands and system safety operation constraints. Fourth, the virtual power plant optimizes the internal resource scheduling plan based on the market clearing results and follows the scheduling instructions issued by the distribution network operator.

[0232] In addition, the internal operation model of virtual power plants plays an important role in the entire power distribution market. Virtual power plants have a variety of flexibility resources and can provide a variety of auxiliary services for the power grid. Virtual power plants should aggregate various internal demand-side and power-side flexibility resources based on their characteristics, consider the coupling relationship between them, and formulate bidding plans to make profits in the power market. After the market is cleared, the virtual power plant will design the dispatch plan of its internal flexibility resources based on the dispatch plan curve and various power commodity prices issued by the distribution network operator, and track the dispatch instructions issued by the distribution network operator.

[0233] (2) Bidding rules for virtual power plants to participate in the electricity market

[0234] Considering that virtual power plants generally contain a large amount of distributed renewable energy, the research work of this invention is carried out on the basis of virtual power plants with high penetration of renewable energy participating in the day-ahead market, intraday demand response market, regulation market and real-time market. The distribution market structure containing virtual power plants is as follows: Figure 3 The bidding rules for virtual power plants are detailed as follows.

[0235] Virtual power plants need to predict their actual power generation 12 to 36 hours in advance and submit bidding information for the next 24 hours to distribution network operators before the day-ahead market closes, i.e., the power generation price-power generation capacity curve. Due to the uncertainty of renewable energy power generation output, it is difficult for virtual power plants to control the deviation between the bid output and the actual output in the day-ahead market. It is assumed that the bidding strategy of the virtual power plant basically does not affect the market price, that is, the virtual power plant is a price taker. In order to ensure that the bid power can win the bid, the virtual power plant can set the bidding price to 0 and optimize its bid power only based on the predicted market settlement price.

[0236] As time goes by, the relevant forecasts will gradually become more accurate, and the virtual power plant can update its predicted power generation. Based on the updated power generation forecast, the virtual power plant can rearrange its dispatch plan and participate in the intraday demand response market to ensure that it is not fined for the deviation between the actual output and the bid output. That is, by purchasing load reduction from the intraday demand response market, the output fluctuations of renewable energy power generation are stabilized. The intraday demand response market closes one hour before the real-time market. Although the output power forecast at this time is more accurate than the day-ahead market forecast, the forecast error is still inevitable. Therefore, the virtual power plant can also use its own energy storage device to stabilize the output fluctuations of renewable energy power generation.

[0237] The regulation market closes one hour before the real-time market opens. During this period, virtual power plants can submit bids for regulation reserves and spinning reserves to distribution network operators. To ensure that the bid reserve capacity can be successful, virtual power plants can set the relevant bids to 0. After the power market transactions end on the day, the actual power generation deviation of the virtual power plant will be settled according to the real-time electricity price.

[0238] Considering that the bidding of virtual power plants in the day-ahead power market and the regulation market is a 1-hour bidding mode, the hourly power generation and hourly power generation of the virtual power plant have the same data. That is, hourly power generation = hourly power generation × 1 hour. Therefore, the present invention uses bidding power generation to study the bidding strategy of virtual power plants.

[0239] (3) Integration characteristics of the electricity market and carbon trading market in which virtual power plants participate

[0240] The operation structure of the virtual power plant in the present invention is as follows: Figure 4 As shown. The power generation unit is mainly composed of wind turbines and gas turbines, the energy storage unit is an electric vehicle, and the demand response unit is a local load and an external demand response provider. In recent years, electric vehicles have developed rapidly. Although electric vehicles are mainly responsible for transportation, they still have energy storage functions. Electric vehicles can be used as an energy storage device to stabilize the output fluctuations of renewable energy power generation and participate in regulating market bidding. Each unit transmits daily operation data to the control center of the virtual power plant through the data layer, and then the control center formulates strategies and dispatches the response capacity of the power generation unit equipment output, energy storage unit and demand response unit in a unified manner according to the current profit and carbon emission targets. Therefore, the power generation units, energy storage units and demand response units in the aggregation area of ​​the virtual power plant are the basis for building an integrated power market and carbon trading market.

[0241] In the carbon trading market, carbon emission rights participate in market transactions as commodities. How to measure carbon emission rights is a key link, and avoiding the repeated use of carbon emission rights is a difficult point in implementation. The integration of the power market and the carbon trading market is not only a circulation and trading center for electric energy, but also a collection and trading center for carbon emission rights. In order to effectively measure carbon emission rights, users trade their carbon emission rights at the same time during the power trading process and obtain corresponding income.

[0242] Virtual power plants participate in integrated transactions in the electricity market and carbon trading market, such as Figure 4 As shown. The process of each unit in the virtual power plant participating in market transactions is also the process of transferring electricity rights and carbon emission rights at the same time, with the following transaction characteristics. First, the virtual power plant needs to purchase corresponding carbon emission rights for the gas-fired units while dispatching gas-fired units to generate electricity. Second, wind turbines do not generate carbon emissions when generating electricity, and virtual power plants can obtain corresponding carbon emission rights. Virtual power plants can allocate carbon emission rights to gas-fired units, or sell them on the carbon trading market to obtain income. Third, the carbon emission rights of energy storage units can be traded simultaneously with electricity.

[0243] 2) Bidding strategies of virtual power plants in the day-ahead market and regulation market

[0244] In order to maximize profits, virtual power plants aggregate wind turbines, electric vehicles and gas turbines to participate in bidding in the day-ahead market and the regulation market. The corresponding bidding strategies are detailed as follows.

[0245] (1) Wind turbines

[0246] First, the income function

[0247] After winning the bid in the day-ahead market, the wind turbine will be dispatched to generate electricity in the real-time market. During the T period, the day-ahead market wind turbine bidding revenue R WT,da , as shown in equation (1).

[0248]

[0249] Where i is the node number of the distribution network; B is the set of distribution network nodes; λ da,e,t is the day-ahead electricity price; is the set of wind turbines connected to node i; P i,wt,t is the day-ahead bidding output of the wind turbine at time t; Δt represents the time interval.

[0250] Second, the cost function

[0251] When the bidding output P WT,t The actual output of wind turbines When there is a deviation, there will be an economic penalty, as shown in equation (2). As the number of dispatchable electric vehicles increases, electric vehicles can be used as energy storage devices for virtual power plants. When the wind turbines are charged, electric vehicles can be charged first, and the remaining electricity can be sold in the real-time market. By controlling the discharge of electric vehicles and purchasing load reduction from the demand response trading market, the deviation between the bid output and the actual output of wind turbines can be effectively reduced, and the output fluctuations of wind turbines can be stabilized.

[0252]

[0253]

[0254]

[0255]

[0256] In the formula, and The actual output is greater than / less than the bid output of the wind turbine; u ec,t 、u bl,t They are the state variables of positive deviation / negative deviation of fan output respectively; and They are the charging and discharging power of electric vehicles to reduce the output deviation of wind turbines; wt,ec is the penalty coefficient for positive deviation of fan output, ω wt,ec <1;ω wt,bl is the penalty coefficient for negative deviation of fan output, ω wt,bl >1; is the actual output of the fan; is the set of electric vehicles connected to node i; γ DR is the set of demand response suppliers; P dr,t is the amount of load reduction purchased from the drth demand response supplier.

[0257] (2) Electric vehicles

[0258] As a controllable load and power source, electric vehicles can provide backup capacity for the distribution network by adjusting their own charging and discharging power while meeting the travel needs of electric vehicles.

[0259] First, the income function

[0260] When electric vehicles participate in the regulation market bidding, the bidding capacity of upward reserve, downward reserve and spinning reserve are expressed as and

[0261] When electric vehicles provide reserve capacity, they cannot predict in advance the output that will be actually dispatched in the real-time market. Therefore, it is necessary to statistically analyze the dispatch ratio of reserve capacity through historical dispatch data. The expected dispatch ratios of electric vehicle up-reserve capacity, down-reserve capacity, and spinning reserve capacity are expressed as and As shown in equation (6-8).

[0262]

[0263]

[0264]

[0265] In the formula, and are the actual dispatch outputs of the upward reserve capacity, downward reserve capacity and spinning reserve capacity of the electric vehicle connected to the node i at time t in the real-time market; and are the upward reserve capacity, downward reserve capacity and spinning reserve capacity of the evth electric vehicle connected to the node i at time t; E(·) represents the expected value.

[0266] The exchange power P between electric vehicles and distribution grid EV,t , as shown in equation (9).

[0267]

[0268]

[0269] In the formula, To meet the charging power required for electric vehicles; and are the total charging and discharging power of the electric vehicle respectively.

[0270] In order to attract users to connect their electric vehicles to charging facilities at designated times and obey the dispatch of virtual power plants, users need to be given discounts on charging costs. ev,sub Charging fees are charged to dispatchable electric vehicles. After winning the bid in the regulation market, the total income of electric vehicles in the regulation market during the T period is R EV,rg , as shown in equation (11).

[0271]

[0272] In the formula, To meet the charging power demand of the evth electric vehicle connected to node i at time t; RD,t ,λ RU,t ,λRR,t They are the capacity prices of downward reserve, upward reserve and spinning reserve in the regulation market respectively.

[0273] Second, the cost function

[0274] The cost of electric vehicles participating in the regulation market bidding consists of two parts, as shown in equation (12).

[0275] C EV =C EV,pur +C EV,los (12)

[0276] First, the cost of electricity purchased by the virtual power plant for charging electric vehicles is C EV,pur , as shown in equation (13).

[0277]

[0278] In the formula, λ rt,e,t Real-time electricity price.

[0279] Secondly, according to relevant experimental data, there is a relationship between the battery loss cost of electric vehicles and the battery purchase cost: the battery loss cost of an electric vehicle with a battery purchase cost of 30,000 yuan is 0.252 yuan / kWh. The dispatchable electric vehicles of the virtual power plant in the present invention use the same type of batteries, and the loss cost of the electric vehicle battery is shown in equation (14).

[0280]

[0281] Where P EV,bat is the rated capacity of the electric vehicle battery; C bat is the purchase cost coefficient of electric vehicle batteries, RMB / kWh. η dc is the discharge efficiency of the electric vehicle battery.

[0282] Although the virtual power plant can determine the number of dispatchable electric vehicles a day in advance, certain reasons may change the number of dispatchable electric vehicles. When electric vehicles cannot be dispatched as planned, the charging and discharging power of other electric vehicles needs to be increased to compensate for the capacity loss. For a certain electric vehicle, when other electric vehicles cannot be dispatched as planned at time t, its charging and discharging power adjustment coefficient is shown in equation (15).

[0283]

[0284] In the formula, To adjust the coefficient, when the total number of electric vehicles reaches 10,000 or more, it can be considered a constant; is the probability that the evth electric vehicle connected to node i at time t cannot be dispatched as planned.

[0285] (3) Gas Generator

[0286] First, the income function

[0287] After winning the bid in the day-ahead market, the gas unit will be dispatched to generate electricity in the real-time market. During the T period, the bidding income of the gas unit in the day-ahead market is R GT,da , as shown in equation (16).

[0288]

[0289] In the formula, is the set of gas units connected to node i; P i,gt,t is the bidding output of the gas generating units in each time period t a few days ago.

[0290] Similar to electric vehicles, when gas units participate in the market bidding, the bidding output of upward reserve, downward reserve and spinning reserve is respectively and express. and They are the actual dispatch value of the upward reserve, the actual dispatch value of the downward reserve and the actual dispatch value of the rotating reserve of the gas unit.

[0291] After the gas unit wins the bid in the regulation market, the total revenue R GT,rg , as shown in equation (17).

[0292]

[0293] Second, the cost function

[0294] The cost of purchasing natural gas from the gas market for the virtual power plant GT,pur , as shown in equation (18).

[0295]

[0296] In the formula, λ gas is the natural gas price; η i,gt,t is the power generation efficiency of the gas unit; LHV is the lower heating value of natural gas, kWh / m 3 .

[0297] 3) Virtual power plant bidding model

[0298] (1) Uncertainty modeling

[0299] The present invention mainly considers three uncertain factors: wind turbine output, local user load demand and market price. Market price includes day-ahead market price, adjustment market price, intraday demand response market price and real-time market price.

[0300] First, fan output

[0301] Although a variety of wind turbine output prediction software has been developed at home and abroad, due to factors such as the error of numerical weather forecasting (NWP), the prediction error of short-term wind turbine output is still 10% to 15%, making it difficult to achieve accurate prediction. In the day-ahead market, the interval constraint method can be used to give the upper and lower limits of wind turbine output to indicate the uncertainty of wind turbine output. The wind turbine bidding output P at time t WT,t The upper and lower limits of the uncertainty parameters P up,WT,t and P low,WT,t The robust optimization method is applicable to optimization problems with uncertain parameters under constraints. At any time t, there is only one uncertain parameter in the upper and lower limit constraints of the wind turbine bidding output, which are P up,WT,t and P low,WT,t The upper and lower limits of wind turbine output follow a normal distribution, as shown in equations (19-20).

[0302]

[0303]

[0304] σ Wind ={σ up,WT,t ,σ low,WT,t} Wind ={μ up,WT,t ,μ low,WT,t} (twenty one)

[0305] In the formula, μ up,WT,t , σ up,WT,t YesP up,WT,t The expected value and standard deviation of low,WT,t , σ low,WT,t YesP low,WT,t The expected value and standard deviation of .

[0306] Chance constraints can be used to convert wind turbine bidding output constraints into inequality constraints, as shown in equations (22-23).

[0307] pr{P WT,t ≤P up,WT,t}≥ε up,WT (twenty two)

[0308] pr{P WT,t ≥P low,WT,t}≥ε low,WT (twenty three)

[0309] In the formula, ε up,WT and ε low,WT The probability of satisfying a chance constraint.

[0310] Second, local user load demand

[0311] Assuming that the virtual power plant needs to meet the load demand of local users, it can be predicted through historical load data, as shown in equation (24).

[0312]

[0313] In the formula, It predicts local user load based on historical load data through the difference autoregressive moving average model; are the upper and lower limits of the prediction error, respectively, and follow a normal distribution.

[0314] Third, market transaction price

[0315] Based on historical data, the day-ahead electricity price, demand response market price, reserve capacity price, and real-time electricity price can be predicted every hour within the range Internal changes. Among them, represents the predicted value of price; Indicates the radius of the price fluctuation range.

[0316] (2) Objective function

[0317] The goal of the virtual power plant bidding strategy in the present invention is to maximize the virtual power plant profit I, as shown in equation (25).

[0318] maxI=R rev -C WT,pl -C EV -C GT,pur -C DR (25)

[0319] R rev =R WT,da +R GT,da +R EV,rg +R GT,rg +R LU +R CT (26)

[0320] In the formula, R rev is the income obtained by the virtual power plant, including: the income of wind turbines and gas turbines in the day-ahead market in equations (1) and (16), the income of electric vehicles and gas turbines in the regulation market in equations (11) and (17), the income from selling electricity to local users in equation (27), and the income from carbon emission rights in the carbon trading market in equation (28). WT,pl It is the economic penalty caused by the deviation between the actual output of the wind turbine in equation (2) and the bid output. EV is the electric vehicle cost in equation (12). C GT,puris the cost of purchasing natural gas for the gas-fired unit in equation (18). DR is the purchase cost of load reduction in the intraday demand response market in equation (30).

[0321] First, income from selling electricity to local users

[0322] Since the load demand of local users shows an elastic response to electricity prices, the relationship between local user electricity consumption and electricity prices can be approximately described by a step price-demand curve, as shown in Figure 7(b). During the T period, the revenue from selling electricity to local users is the product of electricity consumption and electricity price, as shown in equation (27).

[0323]

[0324] In the formula, u lu,t is the ratio of local user electricity price to real-time electricity price, 0≤u i,lu,t ≤1; is the set of local users accessing node i; is the power consumption of the luth user at time t.

[0325] Second, the income from the carbon trading market

[0326] The introduction of the carbon trading mechanism has realized the quantification of carbon emission rights, allowing carbon emission rights to participate in market transactions as commodities. The introduction of carbon trading costs will change the operating costs of gas units. When the carbon emissions of gas units are too high, it is necessary to purchase carbon emission rights from the carbon trading market, which increases the operating costs of gas units. When the excessively high operating costs of gas units affect the profits of virtual power plants, virtual power plants will adjust their operating strategies to limit the bidding output of gas units. Therefore, relying on market mechanisms can limit carbon emissions and improve the energy consumption structure. High-carbon emission enterprises need to pay high costs to purchase carbon emission rights in the carbon trading market, and low-carbon emission enterprises can sell carbon emission rights in the carbon trading market to obtain certain benefits. The introduction of the carbon trading mechanism can promote the transformation of power generation enterprises to green and low-carbon. The income of virtual power plants from selling carbon emission rights in the carbon trading market is shown in equation (28).

[0327]

[0328]

[0329] In the formula, λ c,t is the price coefficient of carbon emission rights at time t, yuan / MW; P i,c,t It is the bidding output of wind turbines and gas turbines connected to node i at time t in the carbon trading market.

[0330] Third, the cost of purchasing load reduction in the demand response market

[0331] In the demand response market, virtual power plants can purchase load reduction from demand response suppliers through bilateral negotiation and centralized bidding, and the purchase cost is shown in equation (30).

[0332]

[0333] In the formula, γ DR is the set of demand response suppliers; C DR,cen is the cost of purchasing load reduction through centralized bidding transactions; C DR,bi It is the cost of purchasing load reduction through bilaterally negotiated transactions.

[0334] Under the centralized bidding mode, the bidding price-bidding quantity curve of the demand response supplier is shown in Figure 8(b). The purchase cost of load reduction under the centralized bidding mode is shown in equation (31).

[0335]

[0336] λ cen,dr,t =θ cen,dr λ rt,e,t (32)

[0337]

[0338]

[0339] In the formula, γ DR is the set of demand response providers; cen,dr,t is the bid price of the drth demand response supplier at time t; P cen,dr,t is the demand response bidding amount corresponding to the demand response bidding price; θ cen,dr is the adjustment coefficient, which can be adjusted by the distribution network operator; Q DR,cen,t is the total demand response provided through the centralized bidding model at time t; is the maximum bid amount of the drth demand response supplier.

[0340] The cost of purchasing load reduction through bilateral negotiation is shown in equation (35).

[0341]

[0342]

[0343] In the formula, λ bi,dr,t Bilaterally negotiated prices for demand response can be agreed upon and determined in advance by both the supply and demand parties; P bi,dr,t is the demand response purchase quantity corresponding to the bilaterally negotiated demand response price; Q DR,Bi,tis the total demand response provided through the bilateral negotiation model at time t.

[0344] The total load reduction purchased by the virtual power plant should meet the demand response capacity DR that the demand response provider can provide. cap Constraint, as shown in equation (37).

[0345]

[0346]

[0347] In the formula, DR cap The demand response capabilities that demand response providers can provide; P dr,t is the amount of load reduction purchased from the drth demand response supplier.

[0348] (3) Constraints

[0349] First, wind turbine operation constraints

[0350] The bid output and actual output of a wind turbine should be between its minimum and maximum output powers, as shown in equations (39-40).

[0351]

[0352]

[0353] In the formula, and are the minimum and maximum output powers of the wind turbine respectively.

[0354] Second, electric vehicle operation constraints

[0355] The states of electric vehicles can be divided into three types: charging, discharging and standby, and they need to satisfy the constraints of equation (41).

[0356]

[0357] In the formula, u dc,t is the discharge state of the electric vehicle. c,t It is the charging status of the electric vehicle.

[0358] The SOC of the electric vehicle at time t is shown in equation (42). In order to extend the battery life of the electric vehicle as much as possible, the battery should be avoided from being fully discharged. Due to reasons such as aging of the electric vehicle battery, there will be a certain amount of rated capacity loss. Therefore, the SOC of the electric vehicle needs to meet the upper and lower capacity constraints, as shown in equation (43).

[0359]

[0360]

[0361] Where: SOC i,ev,t is the SOC value of the evth electric vehicle connected to node i in time period t. SOC i,ev,0 is the initial SOC of the electric vehicle. are the upper and lower limits of the electric vehicle SOC respectively. i,ev,bat is the rated capacity of the battery of the evth electric vehicle connected to node i.

[0362] Electric vehicles also need to meet the maximum charging and discharging power P EV,max Constraints, as shown in equations (44-45).

[0363]

[0364]

[0365] Third, gas unit operation constraints

[0366] The bidding output of the gas unit should meet the upper and lower limits of its output power and the maximum ramp rate constraints, as shown in equations (46-49).

[0367]

[0368]

[0369]

[0370]

[0371] In the formula, is the maximum output power of the gas unit; It is the bidding capacity of gas units to be adjusted down for standby; and They are the maximum upward / downward ramp rates of the gas unit respectively.

[0372] Fourth, line constraints

[0373]

[0374]

[0375] In the formula, is the injected power of node i at time t; P ij,max , P ij,min are the upper and lower limits of the power of line ij respectively; h ij is the power distribution variation coefficient of line ij.

[0376] 4) Robust optimization bidding model for virtual power plants

[0377] The bidding model of the virtual power plant is a mixed integer programming model. This section first converts it into a linear programming problem and then into a robust linear programming model. In order to convert the bidding model of the virtual power plant into a linear programming problem, it is necessary to eliminate the 0-1 variables in the model. Consider introducing decision variables to modify the constraint equations (3-4, 9) to form new constraint equations (52-55).

[0378]

[0379]

[0380]

[0381]

[0382] According to equation (42), there are differences in the charging and discharging efficiency of electric vehicles, which results in different energy losses during the battery charging and discharging process. Equation (42) is converted into a linear function equation (56) to calculate the SOC of electric vehicles.

[0383]

[0384]

[0385] When the electric vehicle is charging, equation (56) is the same as equation (42). When the electric vehicle is discharging, the electric vehicle discharge amount calculated by equation (56) is less than the actual discharge amount due to linearization, and Δerr i,ev,t Indicates the deviation of discharge amount. When the charging and discharging efficiency of electric vehicles is η c =η dc =0.95, Δerr i,ev,t Only accounts for the actual discharge 9.75%. When η c =η dc =1.0, Δerr i,ev,t = 0. Therefore, it is feasible to replace equation (42) with equation (56).

[0386] According to the robust optimization principle, the linear programming model is transformed into a robust optimization bidding model for a virtual power plant, as detailed below.

[0387] max I ROM (58)

[0388]

[0389]

[0390]

[0391]

[0392]

[0393]

[0394]

[0395]

[0396]

[0397]

[0398]

[0399]

[0400]

[0401]

[0402] -y≤P≤y (73)

[0403]

[0404]

[0405]

[0406]

[0407]

[0408]

[0409]

[0410] Where P is the decision variable vector; y is the auxiliary decision variable vector introduced by the dual transformation; z ε is the upper quantile of the standard normal distribution; μ a,W,t and σ a,W,t They are The expected value and standard deviation of It follows a normal distribution.

[0411] In the robust optimization bidding model of virtual power plant bidding, the charging and discharging power constraints of electric vehicles satisfy equations (54) and (55), and the SOC constraints of electric vehicles satisfy equation (56). Equations (77) and (78) are the upper and lower limit constraints of the bidding output of wind turbines, and equations (79) and (80) are the positive and negative deviation constraints of the bidding output of wind turbines. Other constraints are shown in equations (43-51).

[0412] In order to ensure that the constraint violation probability does not exceed κ, the robust control coefficient Γ1 needs to satisfy the constraint equation (81).

[0413]

[0414] In the formula, Φ -1 is the inverse function of the cumulative distribution function of the standard normal distribution; n np is the number of uncertain parameters contained in constraint equation (59).

[0415] 5) As a modeling language for describing and solving large-scale optimization problems, AMPL does not solve optimization problems directly, but obtains the optimal solution by calling appropriate external solvers (such as CPLEX, MINOS, IPOPT, SNOPT, KNITRO, etc.). IBMILOG CPLEX12.2 is a commercial solver for solving quadratic programming and mixed integer linear programming, and has been widely used in the field of optimization. The collaborative bidding strategy of the virtual power plant participating in the electricity market and the carbon trading market of the present invention will be constructed based on the MATLAB R2018b platform using AMPL / CPLEX in an environment where the CPU is Inter(R)Core(TM)i7-8250U and the main frequency is 1.8GHz, and the robust optimization model of the virtual power plant participating in the collaborative bidding in the electricity market and the carbon trading market is solved.

[0416] Best Mode

[0417] The best implementation example takes a virtual power plant demonstration area in a certain place in my country as an example, establishes a robust optimization model for the virtual power plant to participate in the collaborative bidding in the electricity market and the carbon trading market, and obtains the collaborative bidding strategy for the virtual power plant to participate in the electricity market and the carbon trading market proposed by the present invention.

[0418] 1) Basic data

[0419] The present invention uses the virtual power plant demonstration area as an example to carry out case analysis, with hourly time periods, a total of 24 time periods per day and Δt = 1h. The virtual power plant mainly includes two gas units, two wind turbines and 20,000 electric vehicles. Local electricity users are divided into three categories: residential, commercial and industrial. In order to minimize the fines for actual power generation deviations, the virtual power plant will purchase load reductions from the intraday demand response market, and the load reduction will not exceed the demand response capacity DR of the demand response provider. cap .

[0420] The expected value and actual output of wind turbine output constraints, such as Figure 5 As shown. The penalty coefficient for the positive deviation of the wind turbine output is 0.95. The penalty coefficient for the negative deviation of the wind turbine output is 1.05. The power parameters of the gas unit are shown in Table 1. The charging and discharging parameters and dispatchable characteristics of electric vehicles are shown in Tables 2 and 3. The price information of the day-ahead market, regulation market and real-time market is shown in Figure 6(a). The expected value of the real-time market regulation / spinning reserve capacity dispatch ratio is shown in Figure 6(b). The trading price of carbon emission rights is 148.41 yuan / MW. The local user load demand forecast curve is shown in Figure 7(a). The step curve of the relationship between local user load demand and electricity price is shown in Figure 7(b). The forecast deviations of Figures 6(a) and 7(a) are both set to ±15%. In the demand response market, the bilateral negotiation curve for the load reduction provided by the demand response supplier is shown in Figure 8(a); the centralized bidding curve is shown in Figure 8(b). Demand response capability DR of demand response suppliers cap The range of variation is [0,25WM].

[0421] Table 1 Power parameters of gas generator sets

[0422]

[0423] Table 2 Electric vehicle charging and discharging parameters

[0424]

[0425] Table 3 Probability of unplanned departure of electric vehicles and charging power

[0426]

[0427] 2) Analysis of the impact of changes in robust control coefficients on bidding results

[0428] When demand response capability DR cap =10MW, standard deviation of wind turbine output upper / lower limit and actual output σ Wind Take the expected value μ WindWhen the probability of constraint violation is 10%, the impact of the change of robust control coefficient Γ1 on the bidding results of virtual power plants is analyzed. It can be seen from the robust optimization model of virtual power plants that the probability of constraint violation can be adjusted by Γ1. Different constraint violation probabilities correspond to the economic risks of various decisions. The smaller the constraint violation probability, the smaller the economic risk borne by the decision maker. The relevant parameter values ​​corresponding to different constraint violation probabilities and the optimal value of virtual power plant profits are shown in Table 4. Among them, the probability that the opportunity constraint is established ε up,WT , ε low,WT , ε pl,ec , ε pl,bl The values ​​are the same and are both ε.

[0429] Table 4. Related parameter values ​​and optimization results under different constraint violation probabilities

[0430]

[0431] The eight scenarios in Table 4 correspond to different risk preferences of decision makers. From scenario 1 to scenario 8, the robust control coefficient gradually increases, and the probability of constraint violation gradually decreases. At the same time, the probability of the opportunity constraint being established increases, and the risk of the virtual power plant bidding result violating the constraint gradually decreases, indicating that the decision maker's risk aversion continues to increase, and the optimal profit value of the virtual power plant also decreases. In the eight scenarios, the results of the virtual power plant wind turbines and electric vehicles participating in the market bidding are as follows: Fig. 9 As shown in Figure 10.

[0432] Depend on Fig. 9 It can be seen that as the risk aversion of decision makers increases, the bidding output of wind turbines decreases. Correspondingly, the charging and discharging power provided by electric vehicles to smooth the fluctuation of wind turbine output also decreases, as shown in Figure 10(a). When electric vehicles in virtual power plants participate in the regulation market bidding, they can provide regulation reserves and rotating reserves at the same time. Affected by the prices of regulation reserves and rotating reserves, electric vehicles need to coordinate various reserves when participating in the regulation market bidding to maximize profits. Therefore, although the changes in the bidding strategies of electric vehicles in different scenarios can be found in Figure 10, there is no certain trend of change.

[0433] 3) Analysis of the impact of changes in the number of electric vehicles and wind turbine output on bidding results

[0434] In the virtual power plant bidding problem, changes in the number of electric vehicles and wind turbine output will affect the virtual power plant bidding strategy. cap =10MW,σ Wind =0.1μ Wind, robust control coefficient Γ1 = 20, ε = 98.09%, and other parameters are the same as "2) Analysis of the impact of changes in the robust control coefficient on the bidding results", and the impact of changes in the number of electric vehicles and wind turbine output on the bidding results is analyzed. When the number of electric vehicles and wind turbine output changes, the bidding results of the virtual power plant are shown in Table 5. Among them, N EV is the number of electric vehicles, in ten thousand. up,WT,t and μ low,WT,t are the expected values ​​of the upper and lower limits of the wind turbine output respectively. The relative change refers to the change relative to the original situation N EV =2, i.e., scenario 7 in Table 4.

[0435] Table 5. Results of bidding strategies when electric vehicles and wind turbine outputs change

[0436]

[0437] It can be seen from Table 5 that when the relative changes in the number of electric vehicles and the output of wind turbines are the same, the change in the number of electric vehicles has a greater impact on the bidding results of the virtual power plant. That is, the virtual power plant bidding strategy is more sensitive to the number of electric vehicles than to the output of wind turbines. When the number of electric vehicles in the system increases, electric vehicles can not only provide backup for wind turbines, but also participate in the regulation market bidding as energy storage devices to obtain benefits. When the number of wind turbines in the system increases, whether the excess output of wind turbines can participate in the power market bidding depends on whether the energy storage of electric vehicles can provide backup for the fluctuation of wind turbine output, which restricts the impact of wind turbine output on the bidding results of virtual power plants. Therefore, the virtual power plant bidding strategy is relatively insensitive to changes in wind turbine output.

[0438] The results show that the virtual power plant bidding strategy model proposed in this invention can reasonably reflect the economic risks of decision-makers in different scenarios, as well as the coordination relationship between electric vehicles and wind turbines in virtual power plant bidding, and can give corresponding bidding strategies for electric vehicles in the regulation market and wind turbines in the day-ahead market. It is worth noting that when making decisions using the virtual power plant robust optimization bidding model established by this invention, the parameter values ​​of the robust control coefficient and opportunity constraint should be appropriately selected according to the risk preference of the decision-maker.

[0439] 4) Analysis of the impact of wind turbine output standard deviation and demand response capacity changes on bidding results

[0440] When the robust control coefficient Γ1 = 20, ε = 98.09%, the standard deviation of wind turbine output and demand response capacity DR are analyzed. cap The impact of the changes on the bidding results are shown in Figures 11 and 12.

[0441] As shown in Figure 11, the market bidding output of wind turbines today and DRcap It is related to the standard deviation of wind turbine output. When the standard deviation of wind turbine output is the same, DR cap The higher the DR is, the greater the bidding output of the wind turbine will be. As shown in Figures 11 and 12(a), when DR cap At the same time, as the standard deviation of wind turbine output increases, the total amount of electricity bid by wind turbines in the day-ahead market also increases. The reason is that the cost of purchasing load reduction is low. When the demand response capability of the demand response supplier is sufficient to compensate for the output volatility of wind turbines, the virtual power plant will increase the bid output of wind turbines in the day-ahead market in order to effectively utilize wind turbines.

[0442] As shown in Figure 12(b), purchasing load reduction from the demand response market can significantly increase the profit of the virtual power plant. cap After reaching a certain level, as DR cap As the DR increases, the profit of the virtual power plant begins to decrease. At this time, compared with the purchase of load reduction, the profit of the virtual power plant by settling the actual power generation deviation according to the real-time electricity price is greater. cap With the increase of DR cap When DR is smaller, the profit of virtual power plant increases with the decrease of standard deviation of wind turbine output. cap When it is large (such as DR cap =25MWh), the larger the standard deviation of wind turbine output, the more profit the virtual power plant will make.

[0443] When the standard deviation of wind turbine output changes, local users’ electricity income is affected by DR. cap The impact of Fig.13 As shown in the figure, the efficiency of the virtual power plant in the day-ahead market bidding increases with the DR cap The virtual power plant reduces the electricity price of local users to increase the electricity income of local users. The relative changes in the electricity prices of local users are shown in Table 6. At the same time, the access location of local users will also have an impact on the electricity prices of users. Since user 1 is directly connected to the gas unit instead of being powered by the wind turbine, the commercial electricity price of user 1 in Table 6 is the highest.

[0444] Table 6 Relative change ratio of electricity prices for local users

[0445]

[0446] 5) The impact of introducing carbon trading on bidding results

[0447] When the robust control coefficient of demand response capability Γ1=20,ε=98.09%,σ Wind =0.1μ Wind , DRcap =10MW, the impact of introducing carbon trading on the bidding results of virtual power plants is analyzed. Compared with wind turbines, gas turbines have the characteristics of stable output and small fluctuations, and have good economic efficiency. Without considering the cost of carbon emission rights, in order to meet load demand and improve economic efficiency, virtual power plants will fully call on gas turbines to participate in market bidding output. At this time, the optimal profit of the virtual power plant is 316,806.6 yuan. Without considering the cost of carbon emission rights, the output of gas turbines participating in the power market bidding is shown in Figure 14.

[0448] As shown in Figure 14, due to the good economy and output stability of gas-fired units, their 24h bidding output is almost always at full load. However, in the context of the increasingly improved carbon trading market, gas-fired units need to pay high carbon emission rights purchase costs for power generation, which is not conducive to the improvement of the economy of virtual power plants. Therefore, the introduction of the carbon trading mechanism plays a key role in the formulation of virtual power plant bidding strategies and the adjustment of energy structure. When considering the cost of carbon emission rights, the output of gas-fired units participating in the power market bidding is shown in Figure 15.

[0449] From scenario 7 in Table 4, we can see that the optimal profit value of the virtual power plant when considering the carbon emission rights cost is 318,442.7 yuan. Compared with the optimal profit value of the virtual power plant when the carbon emission rights cost is not considered, the profit has increased slightly. By comparing Figures 14 and 15, we can draw the following conclusions.

[0450] First, from 01:00 to 10:00, the bidding output of wind turbines is low, and the virtual power plant calls on gas turbines to meet the power load. From 11:00 to 24:00, the output of wind turbines gradually increases, which can meet the user load demand to a large extent. The introduction of the carbon trading mechanism increases the operating cost of gas turbines. In order to reduce the cost of purchasing carbon emission rights, the virtual power plant reduces the bidding output of gas turbines. The total bidding power generation of gas turbines dropped from 921.12MWh to 758.88MWh.

[0451] Second, the purchase of carbon emission rights has significantly reduced the economic efficiency of gas turbines. In order to increase the optimal profit value of virtual power plants, it is necessary to increase the output of wind turbines. In this example, the output of wind turbines has reached saturation. Therefore, the introduction of the carbon trading mechanism has not had a significant impact on the bidding output of wind turbines.

[0452] Third, the penetration rate of renewable energy in this example is relatively high, and the revenue from the virtual power plant’s sale of carbon emission rights is greater than the cost of purchasing carbon emission rights for gas-fired units, which increases the optimal profit value of the virtual power plant by 1,636.1 yuan.

[0453] Therefore, the introduction of the carbon trading mechanism makes the bidding income of virtual power plants more diversified. By reasonably adjusting the bidding output of wind turbines and gas turbines, virtual power plants can obtain greater economic benefits. At the same time, the introduction of the carbon trading mechanism effectively reduces the emission of pollutants and greenhouse gases in the region, which has both economic and social significance.

[0454] The above reference embodiment provides a detailed description of the collaborative bidding method for a virtual power plant to participate in the electricity market and the carbon trading market. This is illustrative rather than restrictive, and several embodiments may be listed within the limited scope. Therefore, changes and modifications without departing from the overall concept of the present invention should fall within the scope of protection of the present invention.

Claims

1. A collaborative bidding method for a virtual power plant to participate in the electricity market and the carbon trading market, characterized by: It includes four steps: (1) Analysis of the operation mechanism and structure of virtual power plants: Taking a virtual power plant with an operating structure consisting of a power generation unit, an energy storage unit, and a demand response unit as the object, a robust optimization bidding strategy is proposed when the virtual power plant participates in the day-ahead power market, the intraday demand response market, the regulation market, the real-time power market, and the carbon trading market; By analyzing the business model of virtual power plants, the planning of participating in power market bidding, and the integration characteristics of power market and carbon trading market, the diversified income and cost structure of virtual power plants participating in the coordinated bidding of power market and carbon trading market are clarified; (2) Bidding strategies of virtual power plants in the day-ahead market and the regulation market: By establishing a revenue-cost model for wind turbines, electric vehicles, and gas turbines, clarify the bidding strategies of virtual power plants participating in the day-ahead power market and the regulation market; establish a quantitative model for the uncertainty of virtual power plants, and reduce the bidding risks of virtual power plants caused by the uncertainty of renewable energy generation through the demand response market; (3) Virtual power plant bidding model: Introducing the carbon trading mechanism, with the goal of maximizing the profits of virtual power plants, and optimizing the collaborative bidding strategy of virtual power plants participating in the electricity market and the carbon trading market; (4) Robust optimization bidding model for virtual power plants: Considering the different risk preferences of virtual power plant bidding, a robust optimization model for the collaborative bidding strategy of virtual power plants participating in the electricity market and carbon trading market is established to further improve the profits of virtual power plants with different risk preferences participating in the collaborative bidding of the electricity market and carbon trading market. The specific steps of the virtual power plant operation mechanism and structure analysis are as follows: (1) The business model of virtual power plants in the power distribution market The distribution market business model of virtual power plants includes four stages: ① Based on the forecast information of distributed renewable energy, the virtual power plant aggregates its internal resources according to the output range and operating cost to obtain the output curve and cost characteristics of the aggregated flexibility resources; ② The virtual power plant submits bidding prices and quantities of multiple power commodities to the distribution network operator; ③ The distribution network operator collects the bidding information of various operators, considers the bidding output and cost characteristics of various units, proposes market clearing strategies and determines the winning bid price and winning bid quantity of virtual power plants based on the demand for various power commodities and the constraints of system safe operation; ④ The virtual power plant optimizes the internal resource dispatch plan according to the market clearing results and follows the dispatch instructions issued by the distribution network operator. The virtual power plant designs the dispatch plan of its internal flexibility resources according to the dispatch plan curve and various electricity commodity prices issued by the distribution network operator, and tracks the dispatch instructions issued by the distribution network operator; (2) Bidding rules for virtual power plants to participate in the electricity market The bidding rules for virtual power plants are as follows: based on the participation of virtual power plants with high penetration of renewable energy in the day-ahead market, intraday demand response market, regulation market and real-time market, Virtual power plants need to predict their actual power generation 12 to 36 hours in advance and submit bidding information for the next 24 hours to distribution network operators before the day-ahead market closes, i.e., the power generation price-power generation capacity curve; set the bidding price of virtual power plants to 0, and optimize their bidding power only based on the predicted market settlement price; the regulation market is closed one hour before the real-time market opens, during which time virtual power plants submit bids for regulation reserves and spinning reserves to distribution network operators. After the power market transaction ends on the day, the actual power generation deviation of virtual power plants will be settled based on the real-time electricity price. The bidding power generation is used to study the bidding strategy of virtual power plants. The hourly power generation and hourly power generation of virtual power plants have the same data, i.e., hourly power generation = hourly power generation × 1 hour; (3) Integration characteristics of the electricity market and carbon trading market in which virtual power plants participate In the operation structure of the virtual power plant: the power generation unit is mainly composed of wind turbines and gas turbines, the energy storage unit is an electric vehicle, and the demand response unit is a local load and an external demand response provider. Each unit transmits daily operation data to the control center of the virtual power plant through the data layer. The control center formulates strategies and dispatches the output of the power generation unit equipment, the response capacity of the energy storage unit and the demand response unit according to the current profit and carbon emission targets. The power generation units, energy storage units and demand response units in the aggregation area of ​​the virtual power plant are the basis for building an integrated power market and carbon trading market. The process of each unit in the virtual power plant participating in market transactions is also the process of transferring electricity rights and carbon emission rights at the same time. 2) Bidding strategies of virtual power plants in the day-ahead market and regulation market (1) Revenue function of wind turbines After winning the bid in the day-ahead market, when the wind turbine is dispatched to generate electricity in the real-time market, during the T period, the day-ahead market wind turbine bidding revenue R WT,da , as shown in equation (1) Where i is the node number of the distribution network; B is the set of distribution network nodes; λ da,e,t is the day-ahead electricity price; is the set of wind turbines connected to node i; P i,wt,t is the day-ahead bidding output of the wind turbine at time t; Δt represents the time interval; (2) Electric vehicles provide backup capacity for distribution networks First, the income function When electric vehicles participate in the regulation market bidding, the bidding capacity of upward reserve, downward reserve and spinning reserve are expressed as and When electric vehicles provide spare capacity, the dispatch ratio of spare capacity is statistically analyzed through historical dispatch data. The expected dispatch ratios of electric vehicles for upward reserve capacity, downward reserve capacity, and rotating reserve capacity are expressed as and As shown in equation (6-8) In the formula, and are the actual dispatch outputs of the upward reserve capacity, downward reserve capacity and spinning reserve capacity of the electric vehicle connected to the node i at time t in the real-time market; and are the upward reserve capacity, downward reserve capacity and spinning reserve capacity of the ev-th electric vehicle connected to the node i at time t, respectively; E(·) represents the expected value; The exchange power P between electric vehicles and distribution grid EV,t , as shown in equation (9) In the formula, To meet the charging power required for electric vehicles; and are the total charging and discharging power of the electric vehicle respectively; Virtual power plant according to electricity price λ ev,sub Charging fees are charged to dispatchable electric vehicles, and the total income of electric vehicles in the regulation market during period T is R EV,rg , As shown in equation (11), In the formula, To meet the charging power demand of the evth electric vehicle connected to node i at time t; RD,t ,λ RU,t ,λ RR,t They are the capacity prices of downward reserve, upward reserve and spinning reserve in the regulation market; Second, the cost function The cost of electric vehicles participating in the regulation market bidding consists of two parts, as shown in equation (12): C EV =C EV,pur +C EV,los (12) First, the cost of electricity purchased by the virtual power plant for charging electric vehicles is C EV,pur , as shown in equation (13) In the formula, λ rt,e,t Real-time electricity price. Secondly, the dispatchable electric vehicles of the virtual power plant use the same type of batteries, and the loss cost of the electric vehicle battery is shown in equation (14). Where P EV,bat is the rated capacity of the electric vehicle battery; C bat is the purchase cost coefficient of electric vehicle batteries, RMB / kWh. η dc is the discharge efficiency of the electric vehicle battery; (3) Gas Generator First, the income function The gas-fired unit that wins the bid in the day-ahead market will be dispatched to generate electricity in the real-time market. During the T period, the bidding income of the gas-fired unit in the day-ahead market is R GT,da , as shown in equation (16) In the formula, is the set of gas units connected to node i; P i,gt,t is the bid output of the gas generating units at each time period t a day ago; When gas units participate in the market bidding, the bidding output of upward reserve, downward reserve and spinning reserve is and express, and They are the actual dispatch value of the gas unit's upward reserve, the actual dispatch value of the downward reserve, and the actual dispatch value of the rotating reserve. After the gas unit wins the bid in the regulation market, the total revenue R GT,rg , as shown in equation (17) Second, the cost function The cost of purchasing natural gas from the gas market for the virtual power plant GT,pur , as shown in equation (18). In the formula, λ gas is the price of natural gas; η i,gt,t is the power generation efficiency of the gas unit; LHV is the lower heating value of natural gas, kWh / m 3 ; 3) Virtual power plant bidding model (1) Uncertainty modeling Uncertain factors include wind turbine output, local user load demand and market prices, where market prices include day-ahead market prices, adjustment market prices, intraday demand response market prices and real-time market prices. ① Fan output In the day-ahead market, the interval constraint method is used to set the upper and lower limits of the wind turbine output. The wind turbine output P at time t is WT,t The upper and lower limits of the uncertainty parameters P up,WT,t and P low,WT,t Indicates that at any time t, there is only one uncertain parameter in the upper and lower limits of the wind turbine bidding output, which are P up,WT,t and P low,WT,t ; The upper and lower limits of wind turbine output follow a normal distribution, as shown in equation (19-21) s Wind ={σ up,WT,t ,s low,WT,t }m Wind ={μ up,WT,t ,m low,WT,t } (21) In the formula, μ up,WT,t , σ up,WT,t yes Pup,WT,t The expected value and standard deviation of low,WT,t , σ low,WT,t YesP low,WT,t The expected value and standard deviation of The opportunity constraint is used to convert the wind turbine bidding output constraint into an inequality constraint, as shown in equations (22-23): pr{P WT,t ≤P up,WT,t }≥ε up,WT (22) pr{P WT,t ≥P low,WT,t }≥ε low,WT (23) In the formula, ε up,WT and ε low,WT the probability of satisfying the chance constraint; ②Local user load demand The virtual power plant needs to meet the load demand of local users and is predicted based on historical load data, as shown in equation (24): In the formula, It predicts local user load based on historical load data through the difference autoregressive moving average model; are the upper and lower limits of the prediction error, respectively, which follow the normal distribution; ③ Market transaction price According to historical data, the day-ahead electricity price, demand response market price, reserve capacity price and real-time electricity price are all within the range of Internal changes, among which represents the predicted value of price; Indicates the radius of the price fluctuation range; (2) Objective function The goal of the virtual power plant bidding strategy is to maximize the virtual power plant profit I, as shown in equation (25): max I=R rev -C WT,pl -C EV -C GT,pur -C DR (25) R rev =R WT,da +R GT,da +R EV,rg +R GT,rg +R LU +R CT (26) In the formula, R rev is the revenue obtained by the virtual power plant, including: the revenue of wind turbines and gas turbines in the day-ahead market R WT,da and R GT,da , regulating the revenue of electric vehicles and gas units in the market R EV,rg and R GT,rg , the income from selling electricity to local users is R LU , the carbon emission rights income in the carbon trading market R CT ; C WT,pl is the economic penalty caused by the deviation between the actual output of the wind turbine and the bid output, C EV is the cost of electric vehicles, C GT,pur is the cost of purchasing natural gas for the gas-fired unit, C DR is the purchase cost of load reduction in the intraday demand response market. (3) Constraints First, wind turbine operation constraints The bid output and actual output of a wind turbine should be between its minimum and maximum output power, as shown in equations (39-40) In the formula, and are the minimum and maximum output power of the wind turbine respectively; Second, electric vehicle operation constraints The states of electric vehicles include charging, discharging and standby, and need to satisfy the constraints of equation (41) In the formula, u dc,t is the discharge state of the electric vehicle. c,t Is the charging status of the electric vehicle The electric vehicle SOC at time t is shown in equation (42). The electric vehicle SOC needs to satisfy the upper and lower capacity constraints, as shown in equation (43): Where: SOC i,ev,t is the SOC value of the evth electric vehicle connected to node i in period t, SOC i,ev,0 is the initial SOC of the electric vehicle, They are the upper and lower limits of the electric vehicle SOC, P i,ev,bat is the rated capacity of the battery of the evth electric vehicle connected to node i; Electric vehicles also need to meet the maximum charge and discharge power PEV,max constraint, as shown in equations (44-45) Third, gas unit operation constraints The bidding output of the gas unit should meet the upper and lower limits of its output power and the maximum ramp rate constraints, as shown in equations (46-49): In the formula, is the maximum output power of the gas unit; It is the bidding capacity of gas units to be lowered for standby; and They are the maximum upward / downward ramp rates of the gas generator set; Fourth, line constraints In the formula, is the injected power of node i at time t; P ij,max , P ij,min are the upper and lower limits of the power of line ij respectively; h ij is the power distribution variation coefficient of line ij; 4) Robust optimization bidding model for virtual power plants First, the virtual power plant bidding model is transformed into a linear programming problem, and then into a robust linear programming model; a new constraint equation is formed by introducing decision variables (52-55) Convert equation (42) into a linear function equation (56) to calculate the SOC of an electric vehicle, According to the robust optimization principle, the linear programming model is transformed into a robust optimization bidding model for virtual power plants, as detailed below: max I ROM (58) -y≤P≤y (73) Where P is the decision variable vector; y is the auxiliary decision variable vector introduced by the dual transformation; z ε is the upper quantile of the standard normal distribution; μ a,W,t and σ a,W,t They are The expected value and standard deviation of It follows a normal distribution. In the robust optimization bidding model of virtual power plant bidding, the charging and discharging power constraints of electric vehicles satisfy equations (54) and (55), and the SOC constraints of electric vehicles satisfy equation (56); equations (77) and (78) are the upper and lower limit constraints of the bidding output of wind turbines, and equations (79) and (80) are the positive and negative deviation constraints of the bidding output of wind turbines; other constraints are shown in equations (43-51).

2. The collaborative bidding method for virtual power plants to participate in the electricity market and the carbon trading market according to claim 1 is characterized by: The cost function of the wind turbine in the distribution market business model of the virtual power plant will be subject to economic penalties when there is a deviation between the bid output and the actual output of the wind turbine, as shown in equation (2). When the wind turbines are used to charge electric vehicles first, the remaining electricity is sold in the real-time market. By controlling the discharge of electric vehicles and purchasing load reduction from the demand response trading market, the deviation between the bid output and the actual output of wind turbines can be reduced, and the output fluctuation of wind turbines can be stabilized. In the formula, and The actual output is greater than / less than the bid output of the wind turbine; u ec,t 、u bl,t They are the state variables of the positive deviation / negative deviation of the fan output respectively; and They are the charging and discharging power of electric vehicles to reduce the output deviation of wind turbines; wt,ec is the penalty coefficient for positive deviation of fan output, ω wt,ec <1;ω wt,bl is the penalty coefficient for negative deviation of fan output, ω wt,bl >1 is the actual output of the fan; is the set of electric vehicles connected to node i; γ DR is the set of demand response suppliers; P dr,t is the amount of load reduction purchased from the drth demand response supplier.

3. The collaborative bidding method for virtual power plants to participate in the electricity market and the carbon trading market according to claim 1 is characterized by: In the distribution market business model of the virtual power plant, when electric vehicles cannot be dispatched as planned, it is necessary to increase the charging and discharging power of other electric vehicles to compensate for the capacity loss. For a certain electric vehicle, when other electric vehicles cannot be dispatched as planned at time t, its charging and discharging power adjustment coefficient is shown in equation (15). In the formula, To adjust the coefficient, when the total number of electric vehicles reaches 10,000 or more, it is considered a constant; is the probability that the evth electric vehicle connected to node i at time t cannot be dispatched as planned.

4. The collaborative bidding method for virtual power plants to participate in the electricity market and the carbon trading market according to claim 1 is characterized by: The objective function of the virtual power plant bidding model is specifically revenue ① Income from selling electricity to local users The relationship between local user electricity consumption and electricity price: During the T period, the income from selling electricity to local users is the product of electricity consumption and electricity price, as shown in equation (27): In the formula, u lu,t is the ratio of local user electricity price to real-time electricity price, 0≤u i,lu,t ≤1; is the set of local users accessing node i; is the power consumption of the luth user at time t; ② Income from the carbon trading market The income of the virtual power plant from selling carbon emission rights in the carbon trading market is shown in equation (28): In the formula, λ c,t is the price coefficient of carbon emission rights at time t, yuan / MW; P i,c,t is the bidding output of the wind turbine and gas turbine connected to node i at time t in the carbon trading market; ③ Load reduction purchase cost in the demand response market In the demand response market, virtual power plants purchase load reduction through bilateral negotiation and centralized bidding from demand response suppliers. The purchase cost is shown in equation (30): In the formula, γ DR is the set of demand response suppliers; C DR,cen is the cost of purchasing load reduction through centralized bidding transactions; C DR,bi is the cost of purchasing load reduction through bilaterally negotiated transactions; The purchase cost of load reduction in centralized bidding is shown in equation (31): l cen,dr,t =θ cen,dr l rt,e,t (32) In the formula, γ DR is the set of demand response providers; cen,dr,t is the bid price of the drth demand response supplier at time t; P cen,dr,t is the demand response bidding amount corresponding to the demand response bidding price; θ cen,dr is the adjustment coefficient, which is adjusted by the distribution network operator; Q DR,cen,t is the total demand response provided through the centralized bidding model at time t; is the maximum bid amount of the drth demand response supplier; The cost of purchasing load reduction through bilateral negotiation is given by equation (35): In the formula, λ bi,dr,t Bilaterally negotiated prices for demand response, agreed and determined in advance by both the supply and demand parties; P bi,dr,t is the demand response purchase quantity corresponding to the bilaterally negotiated demand response price; Q DR,Bi,t is the total demand response provided through the bilateral negotiation model at time t, The total load reduction purchased by the virtual power plant should meet the demand response capacity DRcap constraint that the demand response provider can provide, as shown in equation (37): In the formula, DR cap The demand response capabilities that demand response providers can provide; P dr,t is the amount of load reduction purchased from the drth demand response supplier.

5. The collaborative bidding method for virtual power plants to participate in the electricity market and the carbon trading market according to claim 1 is characterized by: In the robust optimization bidding model of the virtual power plant, when the electric vehicle is charging, equation (56) is the same as equation (42). When the electric vehicle is discharging, the electric vehicle discharge amount calculated by equation (56) is less than the actual discharge amount, and Δerr is used. i,ev,t Indicates the deviation of discharge capacity. When the charge and discharge efficiency of electric vehicles is η c =η dc =0.95, Δerr i,ev,t Only accounts for the actual discharge 9.75% when η c =η dc =1.0, Δerr i,ev,t =0, replace equation (42) with equation (56).

6. The collaborative bidding method for virtual power plants to participate in the electricity market and the carbon trading market according to claim 1 is characterized by: In the robust optimization bidding model of the virtual power plant, in order to make the constraint violation probability not exceed κ, the robust control coefficient Γ1 needs to satisfy the constraint equation (81) In the formula, Φ -1 is the inverse function of the cumulative distribution function of the standard normal distribution; n np is the number of uncertain parameters contained in constraint equation (59).

7. The collaborative bidding method for virtual power plants to participate in the electricity market and the carbon trading market according to claim 1 is characterized by: The collaborative bidding strategy for the virtual power plant to participate in the electricity market and the carbon trading market will be constructed and solved using AMPL / CPLEX on the MATLAB R2018b platform in an environment where the CPU is Inter(R)Core(TM) i7-8250U with a main frequency of 1.8GHz.

Images