A two-layer optimization method and device for a rural virtual power plant electricity seller system

Abstract

The present invention relates to a two - layer optimization method and device for a rural virtual power plant electricity seller system, belonging to the field of power electronics technology, and solves the problem of how to perform two - layer optimization of VPP - ER, which is an urgent problem to be solved in the transformation of the current new power system. The method includes: establishing multiple unit models based on rural VPP and rural load demands; establishing a collaborative trading mathematical model based on electricity - carbon - green certificate collaborative trading, including a green certificate trading cost model; constructing an upper - layer electricity - carbon - green certificate market collaborative trading optimization model based on the collaborative trading mathematical model, considering the benefits and risk costs of the VPP - ER system participating in the electricity - carbon - green certificate market collaborative trading, and combining the day - ahead electricity price, real - time electricity price, carbon price and green certificate price to allocate the electricity purchase volume of the day - ahead market and the virtual power plant to maximize the weighted benefit; constructing a lower - layer virtual power plant scheduling optimization model based on multiple unit models to minimize the scheduling cost. It can minimize the trading risk and scheduling cost and maximize the comprehensive benefit.

Description

Technical Field

[0001] The present invention relates to the field of power electronics technology, and particularly to a two-layer optimization method and device for a rural virtual power plant electricity retailer system. Background Art

[0002] Reasonably allocating distributed energy in rural areas to meet "self-use" and opening up the "surplus grid connection" channel is the basis for building a new type of intelligent power grid in rural areas. Independent electricity retailers (ERs) are becoming new market players, which enables virtual power plants (VPPs) to participate in power market transactions through signing contracts with electricity retailers. Electricity retailers can also utilize the regulation performance of VPPs to enhance the flexibility of trading schemes, which plays an important role in promoting the grid connection of rural distributed energy generation.

[0003] Promoting the trading of green power certificates can accelerate the research and establishment of a mandatory assessment method for non-hydro renewable energy power generation quotas and improve the renewable energy consumption responsibility weight system. How to consider the two-layer optimization of rural virtual power plant electricity retailers in the electricity-carbon-green certificate coupled market is a key issue that urgently needs to be solved in the transformation of the current new power system.

[0004] Currently, the research on the power purchase and sale trading strategies of independent electricity retailers mainly focuses on the impact of price mechanisms, load deviations, and risk factors on power purchase and sale strategies. Some research results study the combined trading strategies of electricity retailers in the medium- and long-term markets, day-ahead markets, and real-time markets, but fail to consider the VPP as a regulating unit and its impact on the allocation of power purchase quantities. Moreover, the current existing research all takes urban electricity retailers as the main body and fails to consider the resource characteristics of rural areas.

[0005] Secondly, regarding the research on the deeply coupled trading mechanism of the power market, carbon trading market, and green certificate trading market, the existing research analyzes the coupling problems of the electricity, carbon, and green certificate trading markets from a macro level and does not involve the coordinated trading problems of electricity-carbon-green certificate markets of electricity retailers. In fact, with the continuous maturity of the electricity-carbon trading market and the green certificate trading market, as an important new entity in the multi-dimensional coupled market, how electricity retailers participate in electricity-carbon-green certificate coordinated trading is an urgent problem to be solved. In addition, the application of research on the power purchase and sale trading strategies of electricity retailers with VPPs mainly focuses on urban areas and does not involve biomass power generation and waste power generation in rural areas. Summary of the Invention

[0006] In view of the above analysis, the embodiments of the present invention aim to provide a two-layer optimization method and device for a rural virtual power plant electricity retailer system to solve the problem that how to consider the two-layer optimization of rural virtual power plant electricity retailers in the electricity-carbon-green certificate coupled market is an urgent problem to be solved in the transformation of the current new power system.

[0007] On the one hand, an embodiment of the present invention provides a two-layer optimization method for a rural virtual power plant electricity seller system, including: establishing a plurality of unit models based on the rural virtual power plant and the load demand in rural areas, where the plurality of unit modules include a wind power output model, a photovoltaic power output model, a biomass power output model, a hydropower unit power output model, a waste power output model, and a load demand response model; establishing an electricity-carbon-green certificate collaborative trading mathematical model based on electricity-carbon-green certificate collaborative trading, where the electricity-carbon-green certificate collaborative trading mathematical model includes a green certificate trading cost model; constructing an upper-layer electricity-carbon-green certificate market collaborative trading optimization model based on the electricity-carbon-green certificate collaborative trading mathematical model, where the upper-layer electricity-carbon-green certificate market collaborative trading optimization model considers the benefits and risk costs of the rural virtual power plant electricity seller VPP-ER system participating in the electricity-carbon-green certificate market collaborative trading, and combines the day-ahead electricity price, real-time electricity price, carbon price, and green certificate price to allocate the electricity purchase volume of the day-ahead market and the virtual power plant to maximize the weighted benefit; and constructing a lower-layer virtual power plant dispatching optimization model based on the plurality of unit models to minimize the dispatching cost.

[0008] The beneficial effects of the above technical solutions are as follows: The present invention considers the three-dimensional coupled market of electricity-carbon-green certificates and the resource characteristics of rural areas, and proposes a two-layer optimization method for a rural virtual power plant electricity seller considering the electricity-carbon-green certificate coupled market from the perspective of the purchase and sale transactions of the rural virtual power plant electricity seller. A dispatching optimization scheme and an electricity-carbon-green certificate purchase and sale trading strategy for the rural virtual power plant electricity seller under multi-dimensional uncertain factors are established, with the expectation of maximizing the comprehensive benefit while minimizing the trading risk. It is beneficial to convert the value of rural energy resources into economic value and contribute to the construction of a new type of intelligent power grid in rural areas.

[0009] Based on a further improvement of the above method, the electricity-carbon-green certificate collaborative trading includes electricity trading, carbon trading, and green certificate trading. Among them, the electricity trading includes the virtual power plant-electricity seller system participating in the day-ahead trading and real-time trading of the electricity market by purchasing electricity from renewable energy power generators, non-renewable energy power generators, and virtual power plant combinations in different types of markets to maximize the electricity purchase and sale transactions; the carbon trading includes selling the surplus carbon emission allowances in the carbon trading market when there is a surplus of carbon emission allowances; conversely, purchasing carbon emission allowances in the carbon trading market to compensate for the shortage of carbon emission allowances when the carbon emission allowances are insufficient; and the green certificate trading includes selling the surplus green certificates in the green certificate trading market when there is a surplus of green certificates; conversely, purchasing green certificates in the green certificate trading market to compensate for the shortage of green certificates when the green certificates are insufficient.

[0010] Based on a further improvement of the above method, the green certificate trading cost model is expressed as:

[0011]

[0012] Among them, F S is the trading cost of VPP-ER green certificates; k is the quota ratio of renewable energy power generation to the on-grid power, and g VPP,t represents the total output of the rural virtual power plant retailer system VPP-ER, represents the power consumption of the waste incineration power generation WI unit at time t; L Load,t represents the internal load of the rural virtual power plant retailer system at time t; T represents the number of data measurement points. When 1 hour is selected as the time scale, T = 24; the quota ratio k of renewable energy power generation to the on-grid power is calculated by the following formula:

[0013]

[0014]

[0015] Among them, g VPP,t represents the total output of the rural virtual power plant retailer system VPP-ER, and g WPP,t represents the power generation output of the wind power plant WPP at time t, and g PV,t represents the power generation output of the photovoltaic power generation unit PV at time t, and g WI,t represents the power generation output of the waste incineration power generation WI at time t, is the total power consumption for flue gas treatment of the waste incineration power generation WI at time t, and g BPG,t represents the output power of the biogas power generation BPG at time t, and g SHS,t represents the power generation output of the small hydropower station SHS at time t, and △P AL,t represents the scheduling power of the adjustable load AL at time t.

[0016] Based on the further improvement of the above method, the upper-layer electricity-carbon-green certificate market collaborative trading optimization model further includes a carbon emission quota model. Among them, the carbon emission quota model is represented by the following formula:

[0017]

[0018]

[0019]

[0020]

[0021] Among them, E t represents the tradable carbon emission quota of VPP-ER at time t; when E t > 0, it means that VPP-ER sells the remaining carbon emission quota in the carbon trading market and obtains the carbon emission quota selling income. On the contrary, when E t< 0 indicates that VPP-ER needs to purchase the lacking carbon emission allowances, buy insufficient allowances in the carbon trading market, and spend the cost of purchasing carbon emission allowances; represents the carbon emissions that the virtual power plant electricity seller should bear at time t; e Non-RE,t e grid,t e VPP,t respectively represent the carbon emission intensities of non-clean energy power generators, the power grid, and virtual power plant VPP at time t; g Non-RE,t g grid,t represents the electricity purchase amounts of the virtual power plant electricity seller from non-clean energy power generators and the power grid at time t; represents the power generation output of BPG and WI within the virtual power plant VPP called by the virtual power plant electricity seller at time t; represents the initial carbon quota obtained by VPP-ER at time t; η represents the carbon emission allocation coefficient per unit of electricity; g RE,t and respectively represent the electricity amounts purchased by VPP-ER from renewable energy power generators and VPP at time t.

[0022] Based on the further improvement of the above method, the upper-layer electricity-carbon-green certificate market collaborative trading optimization model further includes: an electricity-carbon-green certificate market collaborative trading revenue model:

[0023]

[0024] where, R benefit represents the collaborative trading revenue of VPP-ER; P t se P t pur P t tran represent the electricity selling price, the day-ahead market electricity purchase price, and the real-time market trading price of VPP-ER at time t; when P t tran > 0, it indicates that VPP-ER sells electricity to the distribution network in the real-time stage, otherwise, VPP-ER purchases electricity from the distribution network in the real-time stage; η t represents the day-ahead market electricity purchase ratio of VPP-ER at time t; L t represents the contract electricity quantity signed by VPP-ER at time t; g VPP,t represents the total output of VPP-ER, represents the real-time market electricity purchase quantity of VPP-ER at time t; P t carbon represents the carbon market trading price at time t; C AL,t represents the cost of implementing adjustable load AL by VPP-ER at time t; E tDenote the tradable carbon emission allowances of VPP-ER at time t; F S is the green certificate trading cost of VPP-ER; C VPP,t Denote the cost of dispatching VPP by VPP-ER at time t, including the dispatching costs of different distributed power sources and the cost of user flexible load;

[0025] C AL,t = c AL,t ΔP AL,t ;

[0026] Among them, c AL,t Denote the unit implementation cost of adjustable load AL at time t, and △P AL,t Denote the dispatching power of adjustable load AL at time t;

[0027]

[0028] Among them, P WPP,t and g WPP,t Denote the call price and output of WPP at time t; P PV,t and g PV,t Denote the call price and output of PV at time t; and Denote the power consumption price and power supply price of WI at time t; and Denote the power consumption power and power supply output of WI at time t.

[0029] Based on the further improvement of the above method, the upper-layer electricity-carbon-green certificate market collaborative trading optimization model includes a weighted revenue maximization objective function:

[0030] MaxR VPP = (1 - λ)R benefit - λR risk ;

[0031] Among them, R VPP is the comprehensive weighted revenue of VPP-ER, and R benefit Denote the collaborative trading revenue of VPP-ER; R risk Denote the risk cost brought by the uncertainty of electricity real-time price, carbon price and green certificate trading price to the collaborative trading of VPP-ER in the electricity-carbon-green certificate market; λ represents the risk weight that VPP-ER can accept, and takes any value in the interval [0, 1];

[0032]

[0033] Among them, R m Denote an m-dimensional real number matrix, and the value of m is related to the size of matrix y; is the decision vector, yT = [P t tran , P t carbon , P TCG,t is a multivariate random vector, β represents the threshold value for the decision maker to judge risk, which is determined by the mean of historical data. When it exceeds the mean of historical data, it indicates a certain risk loss; α is the confidence level of the final CVaR, is the completeness of the original scenario set. When the scale of the original scenario set is sufficient to reflect the market price change rule, tends to 1, ρ(y) is the probability density function of the variable y, where, η * is the scenario reduction loss degree between the target scenario set and the original scenario set. The scenario reduction loss degree is calculated by improving the conditional value at risk (CvaR) model:

[0034] Step 1: Calculate the similarity between two scenarios P i and P j using the following formula:

[0035]

[0036] where the scenario set contains N original scenarios and each scenario contains D-dimensional parameters. The i-th scenario is represented as and represent the d-th dimensional parameters in scenarios P i and P j , x sin (P i , P j ) represents the similarity between scenarios P i and scenario P j ; p i and p j are the weights of scenarios P i and scenario P j respectively; ε is a sufficiently small positive number;

[0037] Step 2: Calculate the similarity matrix corresponding to the scenario set using the following formula:

[0038]

[0039] Step 3: Assume that x sin (P i , P j ) is the smallest, then the two scenarios are weighted and combined using the following formula for scenario reduction:

[0040]

[0041] Step 4: Calculate the correlation coefficient of the elements in the \(i\)-th and \(j\)-th dimensions of all scenarios in the \(D\)-dimensional data scenario set \(\Omega\) through the following formula:

[0042]

[0043] where, is the arithmetic mean of the elements in the \(i\)-th and \(j\)-th dimensions of all scenarios in the scenario set \(\Omega\); Calculate the correlation loss function of the new scenario set formed after scenario merging through the following formula:

[0044]

[0045] where, \(\varPhi\) i,j is defined as the new scenario set formed by merging the \(i\)-th scenario and the \(j\)-th scenario in the scenario set \(\Omega\);

[0046] Step 5: Calculate that the final relativity loss matrix can be regarded as the correlation loss matrix of \(\varPhi\) relative to the scenario set \(\Omega\) under each round of scenario reduction through the following formula:

[0047]

[0048] In the formula, represents the number of scenarios in the current sub-scenario set \(\varPhi\);

[0049] Step 6: During the scenario reduction process, when two similar scenarios are merged, a certain amount of data loss will occur. Calculate the scenario reduction loss degree \(\eta\) between the target scenario set \(\varPhi\) and the original scenario set \(\Omega\) through the following formula * :

[0050]

[0051] where, represents the maximum upper triangular element in the difference matrix \(X\) after each round of normalization processing c ; \(K\) represents the number of target scenario sets.

[0052] Based on the further improvement of the above method, the constraint conditions of the weighted revenue maximization objective function include: adjustable load AL output constraint, virtual power plant VPP dispatch output constraint, carbon quota trading constraint, and power purchase margin constraint. Among them, the adjustable load AL output constraint is represented by the following formula:

[0053]

[0054] where, \(\Delta P\) AL,t represents the scheduling power of the adjustable load at time \(t\), represents the maximum output provided by the adjustable load AL; The VPP dispatch output constraint is represented by the following formula:

[0055]

[0056]

[0057] Among them, g VPP,t and g VPP,t-1 represent the total output powers of VPP-ER at time t and time t-1, and represent the maximum and minimum output powers of VPP at time t; and represent the maximum and minimum ramping powers of VPP at time t; The carbon quota trading constraint is represented by the following formula:

[0058]

[0059]

[0060] When , it indicates that all the initial carbon emission quotas obtained by VPP-ER are sold in the carbon trading market, and all the trading electricity of VPP-ER comes from zero-carbon power sources, represents the carbon emissions that the virtual power plant seller should bear at time t; represents the initial carbon quota obtained by VPP-ER at time t; E t represents the tradable carbon emission quota of VPP-ER at time t. The electricity purchase margin constraint is represented by the following formula:

[0061]

[0062] Among them, γ represents the electricity purchase margin of VPP-ER in the day-ahead stage, and its value is higher than η t , which is used to cope with the risk of power shortage or sharp increase in electricity price in the real-time stage, represents the power generation outputs of BPG and WI within VPP called by the virtual power plant seller at time t; L t represents the contract electricity quantity signed by VPP-ER at time t; △P AL,t represents the dispatching power of the adjustable load at time t.

[0063] Based on the further improvement of the above method, the lower-layer virtual power plant dispatching optimization model is represented by the following formula:

[0064]

[0065] In the formula: P WPP,t 、P PV,t 、P BPG,t 、P SHS,t and PWI,t respectively represent the call cost coefficients of WPP, PV, BPG, SHS, and WI at time t; P t se represents the call cost coefficient of WI for handling garbage by consuming electricity within time t, g WPP,t represents the power generation output of the wind power plant WPP at time t, g PV,t represents the power generation output of the photovoltaic power generation unit PV at time t, represents the power generation output of the waste power generation WI at time t, g BPG,t represents the output power of the biogas power generation at time t, g SHS,t represents the power generation output of the small hydropower station SHS at time t, C AL,t represents the cost of implementing AL by VPP-ER at time t; is the total power consumption of the waste power generation WI for handling flue gas at time t.

[0066] Based on the further improvement of the above method, the lower-layer virtual power plant scheduling optimization model further includes the following constraint conditions. Among them, the power supply and demand balance constraint is represented by the following formula:

[0067]

[0068] Among them, is the VPP scheduling plan determined by the upper-layer trading optimization model, △P AL,t represents the scheduling power of the adjustable load at time t; the SHS reservoir water demand, power generation diversion flow, and water abandonment constraints are respectively represented by the following formulas:

[0069]

[0070] Q min ≤Q t ≤Q max ;

[0071] S min ≤S t ≤S max ;

[0072] Among them, is the regulating reservoir water demand at time t - 1, V min is the minimum reservoir storage capacity allowed for reservoir drawdown, V max is the maximum available water volume allowed for the reservoir, q t 、Q t 、S t are respectively the natural inflow, power generation diversion flow, and water abandonment flow of the hydropower station at time t; Q min 、Q max are respectively the minimum and maximum values of the power generation diversion flow of the hydropower station turbine, Smin and S max are the minimum and maximum allowable water discharge amounts of SHS, respectively; the WI operation constraint is expressed by the following formula:

[0073]

[0074]

[0075]

[0076] where Q GR,t and Q GS,t represent the flue gas amounts entering the reaction tower and the gas storage tank at time t, respectively; is the flue gas amount that GS enters GR at time t; represents the maximum gas outlet amount of GS; represents the maximum flow rate of the flue gas pipeline; the carbon emission constraint is expressed by the following formula:

[0077]

[0078] where e BPG,t and e WI,t represent the power generation carbon emission intensities of BPG and WI at time t, and e VPP,t represents the power generation carbon emission intensity of the virtual power plant VPP at time t;

[0079] The green certificate constraint is expressed by the following formula:

[0080]

[0081] where S BPG and S WI represent the green certificate demand amounts of BPG and WI; represents the number of green certificates reserved by the upper-layer trading optimization model of VPP.

[0082] On the other hand, an embodiment of the present invention provides a two - layer optimization device for a rural virtual power plant electricity seller system, including: a unit model construction module, configured to establish a plurality of unit models based on the rural virtual power plant and the load demand in rural areas, wherein the plurality of unit modules include a wind power output model, a photovoltaic power output model, a biomass power output model, a hydropower unit power output model, a waste - to - energy power output model, and a load demand response model; a trading model construction module, configured to establish an electricity - carbon - green certificate collaborative trading mathematical model based on electricity - carbon - green certificate collaborative trading, wherein the electricity - carbon - green certificate collaborative trading mathematical model includes a green certificate trading cost model; an upper - layer electricity - carbon - green certificate market collaborative trading optimization model, configured to construct an upper - layer electricity - carbon - green certificate market collaborative trading optimization model based on the electricity - carbon - green certificate collaborative trading mathematical model, wherein the upper - layer electricity - carbon - green certificate market collaborative trading optimization model considers the benefits and risk costs of the rural virtual power plant electricity seller VPP - ER participating in the electricity - carbon - green certificate market collaborative trading, and combines the day - ahead electricity price, real - time electricity price, carbon price, and green certificate price to allocate the electricity purchase volume of the day - ahead market and the virtual power plant to maximize the weighted benefit; and a lower - layer virtual power plant scheduling optimization model, configured to construct a lower - layer virtual power plant scheduling optimization model based on the plurality of unit models to minimize the scheduling cost.

[0083] Compared with the prior art, the present invention can at least achieve one of the following beneficial effects:

[0084] 1. The present invention considers the three - dimensional coupled electricity - carbon - green certificate market and the resource characteristics of rural areas. Starting from the perspective of the purchase and sale transactions of rural virtual power plant electricity sellers, it proposes a two - layer optimization method for rural virtual power plant electricity sellers considering the electricity - carbon - green certificate coupled market, and establishes a scheduling optimization scheme and an electricity - carbon - green certificate purchase and sale trading strategy for rural virtual power plant electricity sellers under multi - dimensional uncertain factors, aiming to maximize the comprehensive benefit while minimizing the trading risk. It is beneficial to convert the value of rural energy resources into economic value and contribute to the construction of a new type of intelligent rural power grid;

[0085] 2. The present invention takes a new type of independent electricity seller as the main body, couples the electricity trading market, carbon trading market, and green certificate trading market, proposes a collaborative trading mode for the electricity - carbon - green certificate market of rural VPP - ER, constructs an upper - layer electricity - carbon - green certificate market collaborative trading optimization model using the improved CVaR method considering market price uncertainty, and a lower - layer VPP - ER scheduling optimization model, forms a two - layer optimization model for the electricity - carbon - green certificate market collaborative trading of rural VPP - ER, and uses the harmony search algorithm to solve the two - layer model, enabling the electricity seller to maximize the comprehensive benefit while minimizing the trading risk, and better guiding the electricity seller to aggregately utilize rural distributed energy and promoting the transformation of China's power system;

[0086] 3. Based on the traditional CVaR model, the present invention introduces an improved scenario reduction and optimization algorithm to extract a set of typical market scenarios, and uses the loss degree after scenario reduction as the basis for selecting the confidence value; and

[0087] 4. The present invention patent uses the harmony search algorithm to solve the double-layer model. This method is simple and convenient to solve, and the solution efficiency is higher than that of traditional methods. Moreover, the heuristic intelligent algorithm can find the global optimal solution faster.

[0088] In the present invention, the above technical solutions can also be combined with each other to realize more preferred combination schemes. Other features and advantages of the present invention will be described in the subsequent specification. Moreover, some advantages can be made obvious from the specification, or can be understood by implementing the present invention. The objectives and other advantages of the present invention can be realized and obtained from the content specifically pointed out in the specification and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0089] The drawings are only for the purpose of showing specific embodiments, and are not considered as a limitation to the present invention. Throughout the drawings, the same reference signs represent the same components.

[0090] Figure 1 is a flowchart of a double-layer optimization method for a rural virtual power plant-electricity retailer system according to an embodiment of the present invention;

[0091] Figure 2 is a diagram of the operation system of a rural virtual power plant-electricity retailer VPP-ER according to an embodiment of the present invention;

[0092] Figure 3 is a structural diagram of a flue gas treatment system for a waste incineration power plant according to an embodiment of the present invention;

[0093] Figure 4 is a diagram of the framework of the electricity-carbon collaborative trading mode participated by VPP-ER according to an embodiment of the present invention;

[0094] Figure 5 is a block diagram of the construction of a double-layer optimization model according to an embodiment of the present invention; and

[0095] Figure 6 is a flowchart of a double-layer optimization device for a rural virtual power plant-electricity retailer system according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0096] The following specifically describes the preferred embodiments of the present invention with reference to the drawings. The drawings form a part of this application and are used together with the embodiments of the present invention to explain the principles of the present invention, rather than to limit the scope of the present invention.

[0097] Reference Figure 1, a specific embodiment of the present invention discloses a two - layer optimization method for a rural virtual power plant electricity seller system, including: in step S101, multiple unit models are established based on the rural virtual power plant and the load demand in rural areas. Among them, the multiple unit modules include a wind power output model, a photovoltaic power output model, a biomass power generation output model, a hydro - power unit power generation output model, a waste - to - energy power generation output model, and a load demand response model; in step S102, an electricity - carbon - green certificate collaborative trading mathematical model is established based on electricity - carbon - green certificate collaborative trading. Among them, the electricity - carbon - green certificate collaborative trading mathematical model includes a green certificate trading cost model; in step S103, an upper - layer electricity - carbon - green certificate market collaborative trading optimization model is constructed based on the electricity - carbon - green certificate collaborative trading mathematical model. Among them, the upper - layer electricity - carbon - green certificate market collaborative trading optimization model considers the benefits and risk costs of the rural virtual power plant electricity seller VPP - ER system participating in the electricity - carbon - green certificate market collaborative trading, and combines the day - ahead electricity price, real - time electricity price, carbon price, and green certificate price to allocate the electricity purchase volume of the day - ahead market and the virtual power plant to maximize the weighted benefit; and in step S104, a lower - layer virtual power plant scheduling optimization model is constructed based on the multiple unit models to minimize the scheduling cost.

[0098] Compared with the prior art, the two - layer optimization method for the rural virtual power plant electricity seller system provided in this embodiment considers the three - dimensional coupled electricity - carbon - green certificate market and the resource characteristics of rural areas. From the perspective of the purchase and sale transactions of the rural virtual power plant electricity seller, a two - layer optimization method for the rural virtual power plant electricity seller considering the electricity - carbon - green certificate coupled market is proposed, and a scheduling optimization scheme and an electricity - carbon - green certificate purchase and sale trading strategy for the rural virtual power plant electricity seller under multi - dimensional uncertain factors are established, aiming to maximize the comprehensive benefit while minimizing the trading risk. It is beneficial to convert the value of rural energy resources into economic value and contribute to the construction of a new rural intelligent power grid.

[0099] In the following, with reference to Figures 1 to 5 , each step of the two - layer optimization method for the rural virtual power plant electricity seller system according to the embodiment of the present invention will be described in detail.

[0100] In step S101, multiple unit models are established based on the rural virtual power plant and the load demand in rural areas. Among them, the multiple unit modules include a wind power output model, a photovoltaic power output model, a biomass power generation output model, a hydro - power unit power generation output model, a waste - to - energy power generation output model, and a load demand response model.

[0101] a. VPP-ER structure description and internal output unit model modeling. Based on the traditional VPP structure that only considers electricity energy trading, the VPP-ER structure breaks the original barriers of VPP, introduces the carbon trading market and the green certificate trading market, and incorporates biomass fuels such as straw and garbage unique to rural areas into VPP, and signs contracts with electricity sales companies to jointly inject or draw energy into the distribution network.

[0102] a1. VPP-ER structure description. For a large number of distributed power sources such as biomass fuels like straw and garbage, small hydropower stations (SHS), wind power prediction (WPP), and photovoltaic (PV) in rural areas, consider integrating WPP, PV, SHS, biomass power generation (BPG), waste incineration power (WI), gas storage devices (GS), and user flexible loads into VPP, and sign contracts with electricity sales companies to jointly inject or draw electricity into the distribution network (as Figure 2 shown).

[0103] a2. Internal output unit model modeling. The rural VPP-ER includes controllable devices, uncontrollable devices, and the flexible loads of signed users.

[0104] (1) Uncontrollable device output model. For VPP-ER, the uncontrollable devices are WPP and PV. Different from conventional power generation devices, the power generation of WPP and PV is affected by natural wind and solar radiation intensity, and their power generation output is uncertain. Although the wind speed and solar radiation intensity show certain intermittency in the short term and long term, data statistics show that the wind speed and solar radiation intensity approximately follow the Weibull distribution and the Beta distribution. According to the local meteorological prediction data, the relevant parameters of the distribution function can be established, and then the power generation output of WPP and PV can be obtained. The specific modeling of the Weibull distribution function is as follows:

[0105]

[0106] Among them, is the probability density corresponding to the random variable v, v is the random variable, k is the shape parameter, c is the scale parameter. Since the wind speed approximately follows the Weibull distribution, e is a constant term, approximately equal to 2.7. That is, the random variable v in (1a) is the wind speed at any moment, and k and c can be calculated from the mathematical expectation and standard deviation of the wind speed sampling sequence samples. Therefore, the WPP output function is as shown in formula (1b).

[0107]

[0108] Where: g WPP,t represents the power output of the WPP at time t; v t represents the natural wind speed at time t; v in and v out represent the cut-in wind speed and cut-out wind speed respectively; v rated represents the rated wind speed; g WPP,R represents the rated wind speed of the WPP.

[0109] Since the solar radiation intensity approximately follows a Beta distribution, we analyze it according to the general probability density function of the Beta distribution. The general modeling of the probability density of the Beta distribution is shown in formula (2a):

[0110]

[0111] where f(x,a,b) is the probability density corresponding to the random variable x, and a and b are the shape parameters of the Beta distribution. They can be calculated from the expectation and variance of the solar radiation intensity, and β(a,b) is the Beta function.

[0112]

[0113] where f(g PV,t ) is the photovoltaic power generation, r is the solar irradiance in the t period; r max is the maximum solar irradiance in the t period, and a and b can be calculated from the mean and standard deviation of the solar radiation intensity. The expression of the Beta function is shown in formula (2c).

[0114]

[0115] where Γ(a) is the gamma function, and the expression of the Γ function is shown in (2d).

[0116]

[0117] In the formula, e is a constant term, approximately equal to 2.7, and m represents a random variable.

[0118] (2) Controllable device output model. For rural areas, the main controllable units mainly include BPG and SHG. For BPG, there are three ways of direct combustion, biogas, and briquetting. Biomass fuel gasification power generation has better clean characteristics. The relationship between the BPG power generation and fuel consumption is modeled as follows:

[0119]

[0120] Where: g BPG,tDenote the output power of biogas power generation; F p Denote the pressure of biogas power generation; F BPG,t Denote the biogas consumption of BPG for power generation at time t; Is the constant term coefficient; And Are the linear term coefficients of biogas power generation pressure and biogas consumption; Is the quadratic term coefficient.

[0121] For SHS, in the long run, the output of SHS is uncertain. However, when a regulating reservoir is configured for it, SHS can adjust the incoming water volume according to the water required for power generation, and combine with the regulating reservoir water level to determine the hydropower output, making the output controllable. The output of a hydropower station mainly depends on the river runoff and the head height, and its expression is as follows:

[0122]

[0123] In the formula: Denote the available output of SHS at time t; η SHS Denote the power generation efficiency of the hydropower station. ρ represents the acceleration due to gravity at the location of SHS. Q t Denote the diversion flow rate of SHS for power generation at time t; H t Denote the net head of the hydropower station; H t =Z u -Z d , Z u And Z d Respectively denote the water level in front of the SHS dam and the water level at the outlet section of the draft tube.

[0124] The actual output of SHS depends on the output limit of the hydropower unit, specifically as follows:

[0125]

[0126] In the formula: And Respectively denote the minimum and maximum outputs of SHS.

[0127] (3) Waste power generation output model. Waste incineration power generation is mainly achieved through the flue gas treatment system. Due to the constraint of its daily fuel supply, the power generation time arrangement has adjustability. Especially after installing a flue gas storage device, the decoupling of power generation time and flue gas treatment time can be realized, and it can participate in the flexibility trading of VPP-ER. Figure 3 Is the structure of the flue gas treatment system of the waste incineration power plant. According to Figure 3 , The flue gas generated by WI power generation is split into the storage device and the reaction tower. Introduce the flue gas split ratio λ to represent the ratio of the flue gas volume flowing into the reaction tower to the total flue gas volume, and adjust the flue gas volume entering the gas storage tank. The specific mathematical modeling is as follows:

[0128]

[0129] Where: Q t represents the total flue gas volume generated by the power plant of WI at time t; Q GR,t and Q GS,t represent the flue gas volumes entering the reaction tower and the gas storage tank at time t, respectively; g WI,t represents the power generation output of WI at time t; e WI represents the flue gas volume generated per unit power generation of WI.

[0130] During the power generation process of WI, there is also the energy consumption of the air pump, including the energy consumption of the flue gas entering and leaving the gas storage tank and the energy consumption of the flue gas entering the reaction tower from the gas storage tank. Then, the total power consumed by WI for treating flue gas includes two parts: the energy consumption power of the air pump and the power consumed for treating flue gas. The specific modeling is as follows:

[0131]

[0132] Where: is the total power consumption of WI for treating flue gas at time t; w WI is the energy consumption coefficient of WI for treating unit flue gas; is the flue gas volume of GS entering GR at time t; w SP is the energy consumption of the air pump when GS enters the flue gas.

[0133] (4) Load demand response model. Adjustable loads mainly participate in the VPP scheduling optimization through price-based demand response and incentive-based demand response, manifested in the form of interruptible and incentiveable loads. The specific mathematical modeling is as follows:

[0134]

[0135] Where: △P AL,t represents the scheduling power of the adjustable load at time t; and represent the scheduling powers of the adjustable load providing interruptible and incentiveable loads, respectively; η on and η off represent the incentive state and interruptible state of the adjustable load.

[0136] In step S102, an electricity-carbon-green certificate collaborative trading mathematical model is established based on electricity-carbon-green certificate collaborative trading. Among them, the electricity-carbon-green certificate collaborative trading mathematical model includes a green certificate trading cost model.

[0137] b. Rural VPP-ER electricity-carbon-green certificate collaborative trading model. Currently, the traditional electricity-carbon-green certificate collaborative trading model only stays in the research of trading models and the calculation of coupled market benefits, and does not calculate the net carbon emission quotas generated by units, especially biomass fuel gasification power generation in rural areas, resulting in deviations in the collaborative trading model. Therefore, the present invention proposes a rural VPP-ER electricity-carbon-green certificate collaborative trading model, which mainly includes four aspects: electricity-carbon-green certificate collaborative trading mechanism, net carbon emission quota calculation, green certificate market trading model, and electricity-carbon-green certificate collaborative trading benefit calculation.

[0138] b1. Electricity-carbon-green certificate collaborative trading mechanism. The present invention establishes a mathematical model to calculate the carbon emission reduction amount of VPP-ER and the trading model of VPP-ER participating in the green certificate market, and puts forward an urgent need for a multi-dimensional market trading strategy for electricity-carbon-green certificates based on the coupling relationship between electricity price, carbon price and green certificate price. Figure 4 It is the collaborative trading model framework for VPP-ER to participate in the three-dimensional coupled market of electricity-carbon-green certificates.

[0139] According to Figure 4 , the VPP-ER transaction (i.e., electricity-carbon-green certificate collaborative trading) includes three parts: electricity trading, carbon trading and green certificate trading. In electricity trading, VPP-ER participates in the day-ahead trading and real-time trading of the electricity market by purchasing electricity from renewable energy power generators, non-renewable energy power generators and VPP combinations in different types of markets to pursue the goal of maximizing the purchase and sale of electricity transactions. After the VPP-ER dispatching plan is determined, the total carbon emissions and the trading volume in the green certificate market can be calculated. For carbon quotas, carbon trading includes selling the surplus carbon emission quotas in the carbon trading market when there is a surplus of carbon emission quotas; conversely, purchasing carbon emission quotas in the carbon trading market to compensate for the shortage of carbon emission quotas when there is a shortage of carbon emission quotas. For green certificates, green certificate trading includes selling the surplus green certificates in the green certificate trading market when there is a surplus of green certificates; conversely, purchasing green certificates in the green certificate trading market to compensate for the shortage of green certificates when there is a shortage of green certificates. Among them, the coupling relationship between electricity price, carbon price and green certificate price will directly affect the formulation of the dispatching plan.

[0140] b2. Net carbon emission quota calculation. When VPP-ER conducts electricity trading, the main carbon emissions come from three channels: the electricity purchased from non-renewable energy power generators, the electricity purchased from the power grid, and the electricity generated by internal BPG and WI. Then, the carbon emissions from the electricity trading of VPP-ER are calculated as follows:

[0141]

[0142]

[0143] In the formula: Indicates the carbon emissions that the virtual power plant electricity seller should bear at time t; e Non-RE,t e grid,t e VPP,t respectively represent the carbon emission intensities of non-clean energy power generators, the power grid, and the virtual power plant at time t; g Non-RE,t g grid,t represent the electricity purchase volumes of the virtual power plant electricity seller VPP-ER from non-clean energy power generators and the power grid at time t; represents the power generation outputs of the BPG and WI within the VPP called by the virtual power plant electricity seller VPP-ER at time t.

[0144] The method of the present invention selects a free initial carbon emission allocation method based on the power generation volume. The carbon emission allocation coefficient per unit of electricity is mainly determined by the "regional power grid baseline emission factor", and it is approximately considered that its carbon emission quota is proportional to the total electricity sales volume. Then, the carbon emission allocation calculation of VPP-ER is as follows:

[0145]

[0146] In the formula: represents the initial carbon quota obtained by VPP-ER at time t; η represents the carbon emission allocation coefficient per unit of electricity; g RE,t and respectively represent the electricity volumes purchased by VPP-ER from renewable energy power generators and the VPP at time t.

[0147] Formula (8) and formula (9) respectively calculate the actual carbon emissions and the initial carbon emission quota of VPP-ER. Then, the tradable carbon emission quota of VPP-ER is as follows:

[0148]

[0149] In the formula: E t represents the tradable carbon emission quota of VPP-ER at time t; when E t > 0, it means that VPP-ER sells the remaining carbon emission quota in the carbon trading market and obtains the income from selling the carbon emission quota. On the contrary, when E t < 0, it means that VPP-ER needs to purchase the lacking carbon emission quota and buy the insufficient quota in the carbon trading market, and spends the cost of purchasing the carbon emission quota.

[0150] b3. Measurement of the green certificate trading cost or income. When accounting for the green certificates generated by VPP-ER and the green certificates required, the green certificates generated by VPP-ER come from WPP and PV, which are related to the renewable energy power generation volume. The green certificates required by VPP-ER mainly include the flue gas treatment system, the internal load of the system, etc. Its green certificate trading cost model is as follows:

[0151]

[0152] Where: F S is the VPP-ER green certificate trading cost; k is the quota ratio of renewable energy power generation to on-grid power, and g VPP,t represents the total output of the rural virtual power plant electricity seller system VPP-ER, represents the power consumption of the waste incineration power generation WI unit at time t; L Load,t represents the internal load of the rural virtual power plant electricity seller system at time t; T represents the number of data measurement points. When 1 hour is selected as the time scale, T = 24.

[0153] Among them, the quota ratio k of renewable energy power generation to on-grid power is calculated as follows:

[0154]

[0155]

[0156] Where, g VPP,t represents the total output of the rural virtual power plant electricity seller system VPP-ER, g WPP,t represents the power generation output of the wind power unit WPP at time t, g PV,t represents the power generation output of the photovoltaic power generation unit PV at time t, g WI,t represents the power generation output of the waste incineration power generation WI at time t, is the total power consumption for flue gas treatment of the waste incineration power generation WI at time t, g BPG,t represents the output power of the biogas power generation BPG at time t, g SHS,t represents the power generation output of the small hydropower station SHS at time t, △P AL,t represents the scheduling power of the adjustable load AL at time t.

[0157] b4. Measurement of the collaborative trading revenue of the electricity-carbon-green certificate market. The revenue of VPP-ER participating in the collaborative trading of the electricity-carbon-green certificate market includes: electricity market revenue, carbon trading market revenue, and green certificate trading market revenue. Among them, the electricity market revenue is composed of electricity sales revenue, day-ahead market power purchase cost, real-time market trading cost, and VPP dispatching cost. At the same time, to increase the flexibility of its own electricity trading, it is considered that VPP-ER can implement PBDR in the day-ahead market, that is, the subsidy obtained by publishing the load to participate in PBDR, to guide users to adjust the electricity load distribution, so as to optimize the power purchase strategy.

[0158]

[0159] Where: R benefit represents the VPP-ER collaborative trading revenue; P t se 、P tpur and P t tran represent the electricity selling price, the day-ahead market electricity purchase price, and the real-time market trading price of VPP-ER at time t; when P t tran > 0, it means that VPP-ER sells electricity to the distribution network in the real-time stage. Conversely, VPP-ER purchases electricity from the distribution network in the real-time stage. η t represents the day-ahead market electricity purchase ratio of VPP-ER at time t; L t represents the contract electricity quantity signed by VPP-ER at time t; g VPP,t represents the total output of VPP-ER, represents the real-time market electricity purchase quantity of VPP-ER at time t; P t carbon represents the carbon market trading price at time t; C AL,t represents the cost of implementing the adjustable load AL of VPP-ER at time t; E t represents the tradable carbon emission quota of VPP-ER at time t; F S is the green certificate trading cost of VPP-ER; C VPP,t represents the cost of dispatching VPP by VPP-ER at time t, mainly including the dispatching costs of different distributed power sources and the costs of user flexible loads.

[0160] C AL,t = c AL,t ΔP AL,t (14a)

[0161] In the formula: c AL,t represents the unit implementation cost of the adjustable load AL at time t, △P AL,t represents the dispatching power of the adjustable load at time t.

[0162]

[0163] In the formula: P WPP,t and g WPP,t represent the call price and output of WPP at time t; P PV,t and g PV,t represent the call price and output of PV at time t; and represent the power consumption price and power supply price of WI at time t; and represent the power consumption power and power supply output of WI at time t.

[0164] In step S103, an upper-layer electricity-carbon-GC collaborative trading optimization model is constructed based on the electricity-carbon-GC collaborative trading mathematical model. Among them, the upper-layer electricity-carbon-GC collaborative trading optimization model takes into account the revenue and risk costs of the rural virtual power plant electricity seller VPP-ER system participating in the electricity-carbon-GC collaborative trading market. Combining the day-ahead electricity price, real-time electricity price, carbon price, and GC price, it allocates the electricity purchase volume of the day-ahead market and the virtual power plant to maximize the weighted revenue. In step S104, a lower-layer virtual power plant scheduling optimization model is constructed based on multiple unit models to minimize the scheduling cost.

[0165] c. Construct a two-layer optimization model. According to the electricity-carbon-GC collaborative trading mode of VPP-ER, it is necessary to determine its day-ahead electricity purchase ratio, real-time trading electricity volume, and VPP scheduling plan. The present invention designs a two-layer optimization model for electricity-carbon-GC collaborative trading in the market, and transforms the upper-layer trading optimization model into a set of KKT conditional equations, and performs iterative optimization with the lower-layer VPP scheduling model to determine the final trading strategy. Figure 5 Construct a block diagram for the two-layer optimization model of electricity-carbon-GC collaborative trading in the VPP-ER market.

[0166] c1. Upper-layer electricity-carbon-GC collaborative trading optimization model. Compared with the traditional electricity-carbon-GC collaborative trading optimization model, the upper-layer electricity-carbon-GC collaborative trading optimization model in the inventive solution takes into account the impact of the uncertainty of the real-time electricity price, carbon price, and GC trading price on the electricity trading benefit. This model comprehensively considers the revenue and risk costs of VPP-ER participating in the electricity-carbon-GC collaborative trading market, combines information such as the day-ahead electricity price, real-time electricity price, carbon price, and GC price, allocates the electricity purchase volume of the day-ahead market, real-time market, and VPP, and takes the maximum comprehensive weighted revenue as the objective function to determine the trading plan that best suits VPP-ER's own risk tolerance.

[0167] MaxR VPP =(1 - λ)R benefit -λR risk (15a)

[0168] Where R VPP is the comprehensive weighted revenue of VPP-ER, R benefit represents the collaborative trading revenue of VPP-ER; R risk represents the risk cost brought by the uncertainty of the real-time electricity price, carbon price, and GC trading price to VPP-ER's participation in the electricity-carbon-GC collaborative trading market; λ represents the risk weight that VPP-ER can accept, and its value is any value in the interval [0, 1];

[0169] In the cost-loss theory for measuring uncertainty risk, the VaR model is commonly used to express the loss cost f(x,y) brought by the optimized decision result when the confidence level is α, where x and y are the decision variable and the random variable respectively. However, when VPP-ER participates in the electricity-carbon-green certificate market, the uncertainty of the trading price is relatively obvious, and the cost loss beyond the confidence level α cannot be ignored. Therefore, the CVaR model is often established, that is, an additional tail risk loss is added on the basis of VaR to represent the loss cost beyond the confidence level α. CVaR can encompass the quantile and its tail risk compared with VaR, and can better reflect the portfolio risk, becoming a more commonly used risk measurement tool.

[0170] For the VaR model, taking a loss cost function f(x,y) as an example, where x and y are the decision variable and the random variable respectively, under the given confidence level α, the VaR function of the loss cost of f(x,y) is defined as

[0171]

[0172] In the formula, ρ(y) is the probability density function of the variable y; R is the set of real numbers.

[0173] Furthermore, this paper introduces the conditional value at risk method (CVaR) to construct the risk cost function. CVaR refers to measuring the conditional expectation value of the loss exceeding the maximum loss at the confidence level 1-α within a given time period under normal market conditions and a certain confidence level 1-α.

[0174]

[0175] To facilitate the solution of CVaR, the probability density function in formula (15c) is discretized in the solution of the present invention, and the estimated value function of the simplified CVaR is obtained as

[0176]

[0177] In the formula, y n is the nth group of sample data after discretization of y, with a total of N groups; p n is the probability corresponding to each group of sample data. It can be seen from formula (15d) that CVaR consists of 2 parts: VaR and the corresponding tail risk loss. When CVaR reaches the optimal value, its VaR value also reaches the optimal value correspondingly.

[0178] Since the confidence level α in the traditional CVaR is only selected subjectively by humans and lacks a reasonable selection basis. And in the CVaR model, the confidence level essentially reflects the credibility of the market scenario value adopted for the current loss cost. Therefore, the present invention improves CVaR and introduces an improved scenario reduction optimization method, by calculating the sub-scenario set in each round of iteration Similarity matrix And the correlation loss matrix with respect to the relative original scenario set Ω Difference matrix X c , the scenarios to be merged in each round are obtained, and finally the optimal confidence level in CVaR is obtained. The steps to improve the CVaR model are as follows:

[0179] 1) Scenario set Contains N original scenarios and each scenario contains D-dimensional parameters, where the i-th scenario can be expressed as Represents the d-th dimension parameter in scenario P i In, calculate the similarity between two scenarios P i And P j As follows:

[0180]

[0181] In the formula, x sin (P i , P j ) represents the similarity between scenario P i And scenario P j ; p i And p j Are the weights of scenario P i And scenario P j respectively; ε is a sufficiently small positive number.

[0182] 2) Calculate the similarity matrix corresponding to the scenario set Specifically as follows:

[0183]

[0184] 3) Weightedly merge 2 scenarios and perform scenario reduction, specifically as follows:

[0185]

[0186] 4) In the D-dimensional data scenario set Ω, calculate the correlation coefficient of the i-th and j-th dimensional elements of all scenarios, specifically as shown in formula (15h), and calculate the correlation loss function of the new scenario set formed after scenario merging, specifically as shown in formula (15i):

[0187]

[0188] In the formula, Is the arithmetic mean of the i-th and j-th dimensional elements in all scenarios of the scenario set Ω.

[0189]

[0190] In the formula, Φi,j It is defined as the new scene set formed by merging the \(i\)-th scene and the \(j\)-th scene in \(\Omega\).

[0191] 5) Calculating the final relative loss matrix can be regarded as the correlation loss matrix of \(\varPhi\) relative to the scene \(\Omega\) under each round of scene reduction, specifically as follows:

[0192]

[0193] In the formula, represents the number of scenes in the current sub-scene set \(\varPhi\).

[0194] 6) During the scene reduction process, when two similar scenes are merged, a certain amount of data loss will occur. Therefore, the scene reduction loss degree between the target scene set \(\varPhi\) and the original scene set \(\Omega\) is defined and denoted as \(\eta\) * , which is accumulated by the losses generated by each scene reduction, that is

[0195]

[0196] In the formula, represents the maximum upper triangular element in the difference matrix \(X\) after normalization for each round c ; \(K\) represents the number of target scene sets.

[0197] 7) After performing scene reduction to extract the typical scene set, the inevitable price information loss will reduce the confidence level of the expected day-ahead cost. To reasonably reflect the influence of the original scene set and the typical scene set on the confidence level of the day-ahead revenue, the CVaR method is further improved. The confidence level \(\alpha\) of the final CVaR is as follows:

[0198]

[0199] In the formula, is the completeness of the original scene set. When the scale of the original scene set is sufficient to reflect the market price change law, tends to 1.

[0200] Then the specific calculation formula for the risk cost is as follows:

[0201]

[0202] In the formula, \(R\) m represents an \(m\)-dimensional real matrix; the value of \(m\) is related to the matrix size of the matrix \(y\), is the decision vector, is a multivariate random vector, β represents the threshold value for the decision maker to determine risk, which is determined by the mean of historical data, that is, exceeding this value indicates a certain risk loss; ρ(y) is the probability density function of variable y. When the analytical formula of ρ(y) is difficult to determine, an approximate solution algorithm can be constructed, and usually the historical data of y or Monte Carlo simulation sample data are used to estimate the integral term of formula (15d).

[0203] Furthermore, the coordinated trading of electricity-carbon-green certificates needs to comprehensively consider the constraints of purchase-sale balance, power generation capacity of power generators, VPP dispatching output, carbon trading, green certificate trading, and power purchase margin. The specific constraint conditions are as follows:

[0204] (1) Output constraint of adjustable load AL. For AL, the output modes include interruptible and incentive-based modes, which need to meet the minimum and maximum output constraints. In order to avoid the peak-valley inversion of the load curve caused by excessive response, the maximum response amount constraint also needs to be met. The specific constraint conditions are as follows:

[0205]

[0206] In the formula: △P AL,t represents the dispatching power of the adjustable load at time t, represents the maximum output that the adjustable load AL can provide.

[0207] (2) VPP dispatching output constraint. VPP mainly includes distributed units such as WPP, PV, WI, and load demand response. Since both WPP and PV have uncertainties, when VPP-ER formulates the VPP call plan, it needs to consider the VPP output constraint and ramp constraint. The specific constraint conditions are as follows:

[0208]

[0209]

[0210] In the formula: g VPP,t and g VPP,t-1 respectively represent the total output of VPP-ER at time t and time t - 1, and g VPP,t-1 represents the output of VPP-ER at time t - 1; and represent the maximum and minimum output powers of VPP at time t; and represent the maximum and minimum ramp powers of VPP at time t.

[0211] (3) Carbon quota trading constraint. The net carbon emission quota of VPP-ER can be sold or purchased in the carbon trading market, but the maximum sold quota cannot exceed the initial carbon emission quota. The specific constraint conditions are as follows:

[0212]

[0213]

[0214] When it indicates that all the initial carbon emission allowances obtained by VPP-ER are sold in the carbon trading market, that is, all the traded electricity of VPP-ER comes from zero-carbon power sources. Formula (18b) means that the actual carbon emission share of VPP-ER in the whole dispatching period should not exceed the sum of the initial allowance and the purchased allowance. Otherwise, purchasing more allowances will increase the total trading cost of VPP-ER, represents the carbon emissions that the virtual power plant electricity seller should bear at time t; represents the initial carbon allowance obtained by VPP-ER at time t; E t represents the tradable carbon emission allowance of VPP-ER at time t.

[0215] (4) Purchase electricity margin constraint. Due to the uncertainty of real-time electricity price and carbon price, when VPP-ER conducts day-ahead electricity purchase, it will over-purchase some electricity according to the contracted traded electricity volume to increase the margin of electricity trading. The specific constraint conditions are as follows:

[0216]

[0217] In the formula: γ represents the electricity purchase margin of VPP-ER in the day-ahead stage, and its value is higher than η t , which is used to cope with the risk of power shortage or sharp increase in electricity price in the real-time stage, represents the power generation output of BPG and WI within VPP called by the virtual power plant electricity seller at time t; L t represents the contracted electricity volume of VPP-ER at time t; △P AL,t represents the dispatching power of the adjustable load at time t.

[0218] c2. Lower-layer VPP dispatching optimization model. After the upper-layer electricity-carbon-green certificate market collaborative trading optimization model determines the VPP dispatching plan, the lower-layer model needs to integrate resources such as WPP, PV, BPG, SHS, WI and load demand response to meet the dispatching plan. Since the carbon emissions from the power generation of BPG and WI in VPP both belong to VPP-ER, the lower-layer optimization model only considers how to minimize the cost as the optimization goal and collaboratively meet the power balance, carbon balance and green certificate balance. The specific objective function is as follows:

[0219] In the formula: P WPP,t 、P PV,t 、P BPG,t 、P SHS,t and P WI,trespectively represent the call cost coefficients of WPP, PV, BPG, SHS, and WI at time t; P t se represents the call cost coefficient of WI for handling garbage by consuming power within time t, g WPP,t represents the power generation output of the wind power plant WPP at time t, g PV,t represents the power generation output of the photovoltaic power generation unit PV at time t represents the power generation output of the waste power generation WI at time t, g BPG,t represents the output power of the biogas power generation at time t, g SHS,t represents the power generation output of the small hydropower station SHS at time t, C AL,t represents the cost of implementing AL by VPP-ER at time t; is the total power consumption of the waste power generation WI for handling flue gas at time t.

[0220] Furthermore, it is necessary to consider the power supply and demand balance within the VPP, the operation constraints of each distributed power source, and the load demand response constraints, etc. The specific constraint conditions are as follows:

[0221] (1) Power supply and demand balance constraint.

[0222]

[0223] In the formula: is the VPP scheduling plan determined by the upper-layer trading optimization model, △P AL,t represents the scheduling power of the adjustable load at time t.

[0224] (2) SHS operation constraint. Small hydropower is subject to seasonal constraints, with dry seasons and wet seasons. For the dry season, it is necessary to ensure that there is a certain amount of water in the reservoir to meet the night load; for the wet season, the water demand cannot exceed the maximum reservoir capacity, so it is necessary to reasonably regulate the reservoir water volume. Especially in the wet season, if the reservoir water volume has basically reached the maximum capacity, due to the maximum limit of the water intake flow of the water turbine unit, it is also necessary to discharge water to ensure the requirements of the reservoir capacity. Accordingly, the reservoir water demand, water intake flow, and water discharge constraints are as follows:

[0225]

[0226] Q min ≤Q t ≤Q max (22b)

[0227] S min ≤S t ≤S max (22c)

[0228] In the formula: is the required water volume of the regulating reservoir at time t-1; V min is the minimum reservoir storage volume allowed for reservoir drawdown, V max is the maximum available water volume allowed for the reservoir, q t 、Q t 、S t are the natural inflow, power generation diversion flow, and spill flow of the hydropower station at time t respectively; Q min 、Q max are the minimum and maximum values of the power generation diversion flow of the hydropower station turbine respectively, S min 、S max are the minimum and maximum values of the allowable spill water volume of the SHS respectively.

[0229] (3) WI operation constraints. For WI, it mainly includes two parts, GR and GS, involving GR power generation constraints and GS gas storage constraints, which require meeting the maximum operating power constraint and the maximum gas storage volume constraint. The specific constraint conditions are as follows:

[0230]

[0231]

[0232]

[0233] In the formula: Q GR,t and Q GS,t represent the flue gas volumes entering the reaction tower and the gas storage tank at time t respectively; is the flue gas volume that GS enters GR at time t; represents the maximum gas outlet volume of GS; represents the maximum flow rate of the flue gas pipeline.

[0234] (4) Carbon emission constraints. The upper-layer trading optimization model establishes the scheduling plan and the total carbon emission limit of the VPP. Therefore, when optimizing the lower-layer scheduling, the total carbon emissions from BPG and WI cannot exceed the total carbon emission limit of the upper-layer model. The specific constraint conditions are as follows:

[0235]

[0236] In the formula: e BPG,t and e WI,t represent the carbon emission intensities of power generation of BPG and WI at time t, and e VPP,t represents the carbon emission intensity of power generation of the VPP at time t.

[0237] (5) Green certificate constraints. The upper-layer trading optimization model establishes the scheduling plan and the green certificate limit of the VPP. Therefore, when optimizing the lower-layer scheduling, the green certificates used by BPG and WI cannot exceed the internal green certificate limit of the upper-layer model. The specific constraint conditions are as follows:

[0238]

[0239] Where: S BPG and S WI represent the green certificate demand of BPG and WI; represents the number of green certificates reserved by the upper-layer trading optimization model of VPP.

[0240] (6) Other constraints. For VPP, on the basis of considering constraints such as power supply-demand balance, SHS operation, WI operation, and IBDR output, the operation constraints of BPG also need to be considered, including maximum / minimum power constraints, start-stop time constraints, and up / down ramp constraints. At the same time, in order to improve the ability to cope with the uncertainty of WPP and PV, the overall up / down rotational reserve constraint of VPP also needs to be considered when conducting scheduling optimization.

[0241] d. Solving the two-layer optimization model. The purpose of establishing the two-layer model in this method is to obtain the optimal electricity-carbon-green certificate collaborative trading strategy of VPP-ER. Among them, the lower-layer model needs to implement the VPP scheduling plan established by the upper-layer model. When there are large output deviations in the lower-layer WPP and PV and the VPP scheduling plan cannot be achieved, the upper-layer model needs to adjust the trading strategy, which makes the upper-layer trading optimization model and the lower-layer VPP scheduling optimization model have a coupling relationship and are difficult to solve independently. Therefore, this paper selects the harmony search algorithm to solve the two-layer model.

[0242] The harmony search algorithm is based on the similarity between music performance and the optimization process of optimization problems. By combining the principle of music performance with the algorithm, the optimal solution of the optimization problem is found. In the optimization problem, each variable is equivalent to an instrument, and the value range of the variable can also be regarded as the pitch range of the instrument. By adjusting the values of each variable, the optimal solution of the optimization problem is sought. Its calculation steps are as follows:

[0243] Step 1: Determine the following basic parameters of the algorithm according to the actual optimization problem: ① The number of instruments (variables) N; ② The value range of the pitch change of each instrument; ③ The size H HMS , that is, the size of the database (population); ④ The probability H HMCR of taking a harmony variable randomly from the memory bank, that is, the probability of extracting a piece of data from the database; ⑤ The pitch fine-tuning probability P(PAR) and the fine-tuning bandwidth F FW . The probability and amplitude of fine-tuning the extracted harmony; ⑥ The maximum number of creations T max , that is, the maximum value of the number of iterations. If the number of iterations is less than T max , then continue the process; otherwise, end the process.

[0244] Step 2: Initialize the harmony memory library. Randomly generate H HMS harmonies from the solution set of the decision variables of the optimization problem to form the initial memory library as follows:

[0245]

[0246] Step 3: Generate a new harmony. Generate a random number r1 from the interval [0, 1]. If r1 < H HMCR , randomly extract a new harmony from the memory library; otherwise, randomly generate it in the decision variable solution space. If the new harmony comes from the memory library, it needs to be adjusted. Generate a new random number r2 in the interval [0, 1]. If r2 < P(PAR), perform fine-tuning according to the fine-tuning bandwidth F FW .

[0247] Step 4: Update the memory library. Evaluate the new variable. If its evaluation value is better than the worst harmony, replace it; otherwise, keep the content of the original harmony memory library.

[0248] Step 5: Check whether the termination condition is reached. If the number of iterations is less than T max , repeat Steps 3 and 4 until T max is reached.

[0249] Reference Figure 6 , another specific embodiment of the present invention discloses a two-layer optimization device for a rural virtual power plant electricity seller system, including: a unit model construction module 601 for establishing multiple unit models based on the rural virtual power plant and rural area load demands, where the multiple unit modules include a wind power output model, a photovoltaic power output model, a biomass power output model, a hydropower unit power output model, a waste power output model, and a load demand response model; a trading model construction module 602 for establishing an electricity-carbon-green certificate collaborative trading mathematical model based on electricity-carbon-green certificate collaborative trading, where the electricity-carbon-green certificate collaborative trading mathematical model includes a green certificate trading cost model; an upper-layer electricity-carbon-green certificate market collaborative trading optimization model 603 for constructing an upper-layer electricity-carbon-green certificate market collaborative trading optimization model based on the electricity-carbon-green certificate collaborative trading mathematical model, where the upper-layer electricity-carbon-green certificate market collaborative trading optimization model considers the benefits and risk costs of the rural virtual power plant electricity seller VPP-ER participating in the electricity-carbon-green certificate market collaborative trading, combines the day-ahead electricity price, real-time electricity price, carbon price, and green certificate price, and allocates the day-ahead market and virtual power plant electricity purchase amounts to maximize the weighted benefit; and a lower-layer virtual power plant scheduling optimization model 604 for constructing a lower-layer virtual power plant scheduling optimization model based on multiple unit models to minimize the scheduling cost.

[0250] Those skilled in the art can understand that all or part of the processes of implementing the methods of the above embodiments can be completed by instructing relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. Among them, the computer-readable storage medium is a magnetic disk, an optical disk, a read-only memory or a random access memory, etc.

[0251] As described above, only the preferred specific embodiments of the present invention are given, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention.

Claims

1. A two - layer optimization method for a rural virtual power plant electricity seller system, characterized in that, Including: Establishing multiple unit models based on a rural virtual power plant and the load demand in rural areas. Among them, the multiple unit modules include a wind power output model, a photovoltaic power output model, a biomass power output model, a hydropower unit power output model, a waste power output model, and a load demand response model; Establishing a mathematical model for electricity-carbon-green certificate collaborative trading based on electricity-carbon-green certificate collaborative trading. Among them, the mathematical model for electricity-carbon-green certificate collaborative trading includes a green certificate trading cost model; Constructing an upper-layer electricity-carbon-green certificate market collaborative trading optimization model based on the mathematical model for electricity-carbon-green certificate collaborative trading. Among them, the upper-layer electricity-carbon-green certificate market collaborative trading optimization model considers the revenue and risk costs of the rural virtual power plant electricity seller VPP-ER system participating in the electricity-carbon-green certificate market collaborative trading, combines the day-ahead electricity price, real-time electricity price, carbon price, and green certificate price, and allocates the electricity purchase volume in the day-ahead market and the virtual power plant to maximize the weighted revenue; and Constructing a lower-layer virtual power plant dispatching optimization model based on the multiple unit models to minimize the dispatching cost. Among them, the upper-layer electricity-carbon-green certificate market collaborative trading optimization model includes a weighted revenue maximization objective function: MaxR VPP = (1 - λ)R benefit - λR risk ; Among them, R VPP is the comprehensive weighted income of VPP-ER, and R benefit represents the collaborative trading income of VPP-ER; R risk represents the risk cost brought by the uncertainties of the real-time electricity price, carbon price and green certificate trading price to the collaborative trading of VPP-ER in the electricity-carbon-green certificate market; λ represents the risk weight that VPP-ER can accept, and its value is any value in the interval [0,1]; Among them, R m represents an m-dimensional real matrix; the value of m is related to the matrix size of matrix y, is the decision vector, y is a multivariate random vector, β represents the threshold value for the decision maker to judge risk, which is determined by the mean of historical data, and indicates a certain risk loss when exceeding the mean of historical data; α is the confidence level of the final conditional value at risk CVaR, is the completeness of the original scenario set. When the scale of the original scenario set is sufficient to reflect the law of market price changes, tends to 1, ρ(y) is the probability density function of variable y, where η * is the scenario reduction loss degree between the target scenario set and the original scenario set, P t tran represents the real-time market trading price of VPP-ER at time t; when P t tran >0, it means that VPP-ER sells electricity to the distribution network in the real-time stage. On the contrary, VPP-ER purchases electricity from the distribution network in the real-time stage; P t carbon represents the carbon market trading price at time t; η t represents the proportion of electricity purchased by VPP-ER in the day-ahead market at time t; represents the real-time market electricity purchase volume of VPP-ER at time t, g VPP,t represents the total output of the rural virtual power plant electricity seller system VPP-ER at time t.

2. The double-layer optimization method of the rural virtual power plant electricity seller system according to claim 1, wherein, The electricity-carbon-green certificate collaborative trading includes electricity trading, carbon trading, and green certificate trading, where The electricity trading includes the virtual power plant-electricity seller system participating in the day-ahead trading and real-time trading of the electricity market by purchasing electricity from renewable energy power generators, non-renewable energy power generators, and virtual power plant combinations in different types of markets to maximize the electricity purchase and sale transactions; The carbon trading includes selling the surplus carbon emission allowances in the carbon trading market when there is a surplus of carbon emission allowances; conversely, purchasing carbon emission allowances in the carbon trading market to compensate for the shortage of carbon emission allowances when there is a shortage of carbon emission allowances; and The green certificate trading includes selling the surplus green certificates in the green certificate trading market when there is a surplus of green certificates; conversely, purchasing green certificates in the green certificate trading market to compensate for the shortage of green certificates when there is a shortage of green certificates.

3. The double-layer optimization method of the rural virtual power plant electricity seller system according to claim 2, wherein The green certificate trading cost model is expressed as: Among them, F S is the trading cost of VPP-ER green certificates; k is the quota ratio of renewable energy power generation to on-grid power; represents the power consumption of the waste-to-power WI unit at time t; L Load,t represents the internal load of the rural virtual power plant retailer system at time t; T represents the number of data measurement points. When 1 hour is selected as the time scale, T = 24; Calculating the quota ratio k of renewable energy power generation to the grid-connected power by the following formula: Among them, g WPP,t represents the power generation output of the wind power plant WPP at time t, and g PV,t represents the power generation output of the photovoltaic power generation unit PV at time t. represents the power generation output of the waste power generation WI at time t. is the total power consumption for flue gas treatment of the waste power generation WI at time t, and g BPG,t represents the output power of the biogas power generation BPG at time t, and g SHS,t represents the power generation output of the small hydropower station SHS at time t, and △P AL,t represents the scheduling power of the adjustable load AL at time t.

4. The double-layer optimization method of the rural virtual power plant electricity seller system according to claim 3, characterized in that The upper-layer electricity-carbon-green certificate market collaborative trading optimization model also includes a carbon emission allowance model, where The carbon emission allowance model is expressed by the following formula: Among them, E t represents the tradable carbon emission quota of VPP-ER at time t; when E t > 0, it means that VPP-ER sells the remaining carbon emission quotas in the carbon trading market and obtains the income from selling carbon emission quotas. On the contrary, when E t < 0, it means that VPP-ER needs to purchase the lacking carbon emission quotas and buy the insufficient quotas in the carbon trading market, and spends the cost of purchasing carbon emission quotas; represents the carbon emissions that the virtual power plant electricity seller should bear at time t; e Non-RE,t , e grid,t , e VPP,t respectively represent the carbon emission intensities of non-clean energy power generators, the power grid, and the virtual power plant VPP at time t; g Non-RE,t , and g grid,t represent the electricity purchase amounts of the virtual power plant electricity seller from non-clean energy power generators and the power grid at time t; represents the power generation outputs of BPG and WI within the virtual power plant VPP called by the virtual power plant electricity seller at time t; represents the initial carbon quota obtained by VPP-ER at time t; η represents the carbon emission allocation coefficient per unit of electricity; g RE,t and respectively represent the electricity amounts purchased by VPP-ER from renewable energy power generators and VPP at time t.

5. The double-layer optimization method of the rural virtual power plant electricity seller system according to claim 3, characterized in that, The upper-layer electricity-carbon-green certificate market collaborative trading optimization model also includes: an electricity-carbon-green certificate market collaborative trading revenue model: Among them, R benefit represents the VPP-ER collaborative trading revenue; P t se and P t pur represent the electricity selling price and the day-ahead market electricity purchase price of VPP-ER at time t; L t represents the contract electricity quantity signed by VPP-ER at time t; C AL,t represents the cost of implementing the adjustable load AL by VPP-ER at time t; E t represents the tradable carbon emission quota of VPP-ER at time t; F S is the VPP-ER green certificate trading cost; C VPP,t represents the cost of dispatching VPP by VPP-ER at time t, including the dispatching costs of different distributed power sources and the costs of user flexible loads: C AL,t = c AL,t ΔP AL,t ; Among them, c AL,t represents the implementation cost per unit of the adjustable load AL at time t, and △P AL,t represents the scheduling power of the adjustable load AL at time t; Among them, P WPP,t and g WPP,t represent the call price and output of WPP at time t; P PV,t and g PV,t represent the call price and output of PV at time t; and represent the power consumption price and power supply price of WI at time t; and represent the power consumption power and power supply output of WI at time t.

6. The double-layer optimization method for the rural virtual power plant electricity seller system according to claim 5, characterized in that, Calculating the scenario reduction loss degree by improving the conditional value at risk CvaR model: Step 1: Calculate the similarity between two scenarios P i and P j ; Step 2: Calculate the similarity matrix corresponding to the scenario set through the following formula : Among them, x sim (P i , P j ) represents the similarity between scenario P i and scenario P j ; Step 3: Set x sim (P i ,P j ) is minimized, then the two scenarios are weighted and combined through the following formula for scenario reduction: Among them, the scene set contains N original scenes, and each scene contains D-dimensional parameters. The i-th scene is represented as and represents the d-th dimensional parameter in scene P i and P j The weights of the d-th dimensional parameter in P i and p j are the weights of scene P i and scene P j respectively; Step 4: Calculating the correlation coefficient between the i-th and j-th elements of all scenarios in the D-dimensional data scenario set Ω by the following formula: Among them, is the arithmetic mean of the elements in the i-th and j-th dimensions in all scenarios in the scenario set Ω; the correlation loss function of the new scenario set formed after scenario merging is calculated by the following formula: Among them, Φ i,j is defined as the new scenario set formed by merging the i-th scenario and the j-th scenario in the scenario set Ω; Step 5: Calculating that the final relative loss matrix can be regarded as the correlation loss matrix of Φ relative to the scenario Ω under each round of scenario reduction by the following formula: In the formula, represents the number of scenarios in the current sub-scenario set Φ; Step 6: During the scenario reduction process, certain data loss will occur when two similar scenarios are merged. Calculate the scenario reduction loss degree η between the target scenario set Φ and the original scenario set Ω through the following formula * : Among them, represents the largest upper triangular element in the differential matrix X after each round of normalization processing c ; K represents the number of target scenario sets.

7. The double-layer optimization method of the rural virtual power plant electricity seller system according to claim 6, characterized in that The constraint conditions of the weighted revenue maximization objective function include: adjustable load AL output constraint, virtual power plant VPP dispatching output constraint, carbon quota trading constraint, and electricity purchase volume margin constraint, where The adjustable load AL output constraint is expressed by the following formula: Among them, △P AL,t represents the scheduling power of the adjustable load at time t, and represents the maximum output provided by the adjustable load AL; The VPP dispatching output constraint is expressed by the following formula: where, g VPP,t and g VPP,t-1 represent the total output powers of VPP-ER at time t and time t-1, and represent the maximum and minimum output powers of VPP at time t; and represent the maximum and minimum ramp powers of VPP at time t; The carbon quota trading constraint is expressed by the following formula: When it indicates that all the initial carbon emission allowances obtained by VPP-ER are sold in the carbon trading market, and all the traded electricity of VPP-ER comes from zero-carbon power sources, represents the carbon emissions that the virtual power plant electricity seller should bear at time t; represents the initial carbon quota obtained by VPP-ER at time t; E t represents the tradable carbon emission quota of VPP-ER at time t; The electricity purchase margin constraint is expressed by the following formula: Among them, γ represents the power purchase margin of VPP-ER in the day-ahead stage, and its value is higher than η t , which is used to cope with the risk of power shortage or sharp increase in electricity price in the real-time stage, represents the power generation output of the BPG and WI within the VPP called by the virtual power plant electricity seller at time t; L t represents the contract electricity quantity signed by VPP-ER at time t; △P AL,t represents the dispatching power of the adjustable load at time t.

8. The double-layer optimization method of the rural virtual power plant electricity seller system according to claim 5, characterized in that The lower-level virtual power plant scheduling optimization model is expressed by the following formula: Where: P WPP,t , P PV,t , P BPG,t , P SHS,t and P WI,t respectively represent the calling cost coefficients of WPP, PV, BPG, SHS, and WI at time t; P t se represents the call cost coefficient for WI to process garbage with the power consumption within time t, g WPP,t represents the power generation output of the wind power plant WPP at time t, g PV,t represents the power generation output of the photovoltaic power generation unit PV at time t represents the power generation output of the waste incineration power generation WI at time t, g BPG,t represents the output power of the biogas power generation at time t, g SHS,t represents the power generation output of the small hydropower station SHS at time t, C AL,t represents the cost of implementing AL by VPP-ER at time t; is the total power consumption for the waste incineration power generation WI to process flue gas at time t 9. The double-layer optimization method of the rural virtual power plant electricity seller system according to claim 8, characterized in that, The lower-level virtual power plant scheduling optimization model further includes the following constraint conditions: The power supply-demand balance constraint is expressed by the following formula: Among them, is the VPP scheduling plan determined for the upper-layer transaction optimization model, and △P AL,t represents the scheduling power of the adjustable load at time t; The water demand, power generation diversion flow, and water abandonment constraints of the SHS reservoir are respectively expressed by the following formula: Q min ≤Q t ≤Q max ; S min ≤ S t ≤ S max ; Among them, V t-1 is the required water volume of the regulating reservoir at time t - 1, V min is the minimum reservoir storage volume allowed for reservoir drawdown, V max is the maximum available water volume allowed for the reservoir, q t and Q t and S t are respectively the natural inflow, power generation diversion flow, and water discharge flow of the hydropower station at time t; Q min and Q max are respectively the minimum and maximum values of the power generation diversion flow of the hydropower station's water turbine, S min and S max are respectively the minimum and maximum values of the allowable water discharge of the SHS; The WI operation constraint is expressed by the following formula: Among them, Q GR,t and Q GS,t respectively represent the amount of flue gas entering the reaction tower and the gas storage tank at time t; is the amount of flue gas that GS enters GR at time t; represents the maximum gas outlet volume of GS; represents the maximum flow rate of the flue gas pipeline; The carbon emission constraint is expressed by the following formula: Among them, e BPG,t and e WI,t represent the power generation carbon emission intensity of BPG and WI at time t, and e VPP,t represents the power generation carbon emission intensity of the virtual power plant VPP at time t; The green certificate constraint is expressed by the following formula: Among them, S BPG and S WI represent the green certificate demand of BPG and WI; represents the number of green certificates reserved by the upper-layer trading optimization model of VPP.

10. A two - layer optimization device for a rural virtual power plant electricity seller system, characterized in that, Including: The unit model construction module is used to establish multiple unit models based on the rural virtual power plant and the rural area load demand. Among them, the multiple unit modules include a wind power output model, a photovoltaic power output model, a biomass power output model, a hydropower unit power output model, a waste power output model, and a load demand response model; The trading model construction module is used to establish a collaborative trading mathematical model based on the electricity-carbon-green certificate collaborative trading. Among them, the electricity-carbon-green certificate collaborative trading mathematical model includes a green certificate trading cost model; The upper-level electricity-carbon-green certificate market collaborative trading optimization model is used to construct the upper-level electricity-carbon-green certificate market collaborative trading optimization model based on the electricity-carbon-green certificate collaborative trading mathematical model. Among them, the upper-level electricity-carbon-green certificate market collaborative trading optimization model considers the benefits and risk costs of the rural virtual power plant seller VPP-ER participating in the electricity-carbon-green certificate market collaborative trading, and combines the day-ahead electricity price, real-time electricity price, carbon price, and green certificate price to allocate the electricity purchase volume in the day-ahead market and the virtual power plant to maximize the weighted benefit; and The lower-level virtual power plant scheduling optimization model is used to construct the lower-level virtual power plant scheduling optimization model based on the multiple unit models to minimize the scheduling cost. Among them, the upper-level electricity-carbon-green certificate market collaborative trading optimization model includes a weighted benefit maximization objective function: MaxR VPP =(1 - λ)R benefit - λR risk ; Among them, R VPP is the comprehensive weighted income of VPP-ER, and R benefit represents the collaborative trading income of VPP-ER; R risk represents the risk cost brought by the uncertainties of the real-time electricity price, carbon price, and green certificate trading price to the collaborative trading of VPP-ER in the electricity-carbon-green certificate market; λ represents the risk weight that VPP-ER can accept, and its value is any value in the interval [0, 1]; Among them, R m represents an m-dimensional real matrix; the value of m is related to the matrix size of matrix y, is the decision vector, y is a multivariate random vector, β represents the threshold value for the decision maker to judge risk, which is determined by the mean of historical data, and indicates a certain risk loss when exceeding the mean of historical data; α is the confidence level of the final conditional value at risk CVaR, is the completeness of the original scenario set. When the scale of the original scenario set is sufficient to reflect the market price change law, tends to 1, ρ(y) is the probability density function of variable y, where η * is the scenario reduction loss degree between the target scenario set and the original scenario set, P t tran represents the real-time market trading price of VPP-ER at time t; when P t tran >0, it means that VPP-ER sells electricity to the distribution network in the real-time stage. Conversely, VPP-ER purchases electricity from the distribution network in the real-time stage; P t carbon represents the carbon market trading price at time t; η t represents the proportion of electricity purchased by VPP-ER in the day-ahead market at time t; represents the real-time market electricity purchase volume of VPP-ER at time t.

Images