Green electricity bidding decision-making model construction method and system considering prediction uncertainty

Abstract

The invention discloses a green electricity bidding decision model construction method and system considering prediction uncertainty. The method comprises the following steps: constructing a green electricity user decision model, a green electricity generator decision model and a green electricity transaction clearing model considering prediction green certificate and carbon quota price uncertainty; and converting the green electricity transaction clearing model into a green electricity user decision model based on the primitive-dual theorem, obtaining a multi-MPEC model according to the constraint conditions of the green electricity generator decision model, carrying out linear processing on the MPEC model, solving the multi-MPEC model by adopting an improved diagonalization algorithm, and finally obtaining a green electricity transaction strategy of the user. According to the method, the relative importance degree of different green evidence and carbon quota price prediction uncertainty variables can be objectively reflected, and the risk of prediction uncertainty is effectively quantified; and the diagonalization algorithm is improved, so that the market equalization problem of the robust optimization model can be efficiently solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power energy optimization allocation, and more specifically, relates to a method and system for constructing a green power bidding decision-making model considering prediction uncertainty. Background Technique

[0002] To promote the realization of "carbon peak and carbon neutrality", China has accelerated the construction of the power market and the carbon market, and incorporated the indirect carbon emissions generated by power consumption on the user side into the carbon market compliance assessment. To further promote energy transformation and green consumption, a medium- and long-term power trading variety with green power products as the subject matter, namely green power trading (hereinafter referred to as "green power trading"), was officially launched in 2021. As a "certificate-electricity integration" form of green power consumption, it naturally has the characteristic of zero carbon emissions. Therefore, some carbon markets have started to pilot relevant policies that do not include green power trading electricity in the indirect carbon emissions accounting. Under the above background, how to optimize its own bidding strategy to obtain higher benefits has become a key issue faced by the current user side when participating in green power trading.

[0003] CN117634831A discloses a method and system for the collaborative equilibrium analysis of power generation enterprises participating in the electricity-carbon multi-market. This invention takes into account the coupling mechanism between the power market and the carbon market, considers the interaction relationship between the carbon market and the power market at different time scales, constructs a two-layer decision-making model for the electricity-carbon multi-market, and converts the two-layer decision-making model for the electricity-carbon multi-market into a single-layer equilibrium constraint programming model to make the single-layer equilibrium constraint programming model more in line with the actual market operation rules and accurately reflect the market operation rules and actual situations. However, this invention only considers the ordinary power market and the carbon market, ignoring the impact of green power policies and the uncertainty of green certificate and carbon quota price forecasts on trading decisions.

[0004] CN117094745A discloses an integrated energy system optimization control method and device based on IGDT-utility entropy. This invention introduces the carbon trading volume into the energy hub model to establish a carbon-energy coupling model; considers the capacity configuration of the IES for optimal operation, takes the minimum total planning cost as the optimization goal, and establishes a deterministic IES optimization control model; establishes an IES optimization control model based on IGDT, uses the uncertain sets of carbon trading prices and energy prices to describe uncertainties; establishes an IES optimization control model based on utility entropy, and obtains the wind-solar scene set through Latin hypercube sampling and K-means clustering method; establishes an IES optimization control model based on IGDT-utility entropy, and analyzes the IES optimization control model to obtain the optimal control scheme and robustness coefficient under a given total cost budget. However, this invention uses the traditional equal-weight IGDT method, which ignores the relative importance of uncertain factors when dealing with decision risk control. Summary of the Invention

[0005] To solve the following problems in the prior art:

[0006] (1) The impact of uncertainty in green certificate and carbon quota price forecasts on green electricity trading decisions is not considered;

[0007] (2) The traditional equal-weighted IGDT method ignores the importance of uncertain factors;

[0008] (3) The market equilibrium problem of the robust optimization model based on information decision theory is difficult to solve;

[0009] The present invention proposes a method for constructing a green electricity bidding decision model taking into account prediction uncertainty.

[0010] The present invention adopts the following technical solution.

[0011] A first aspect of the present invention provides a method for constructing a green power bidding decision model considering prediction uncertainty, comprising:

[0012] Step 1: Considering the uncertainty of green certificate and carbon quota price forecast, a green electricity user decision model is constructed;

[0013] Step 2: Construct a green power generator decision model;

[0014] Step 3: Construct a green electricity trading clearing model;

[0015] Step 4: Convert the constructed model into a multi-MPEC model based on the primal-dual theorem;

[0016] Step 5: Linearize the multi-MPEC model and solve the MPEC model using the improved diagonalization algorithm.

[0017] Preferably, in step 1, the green electricity user decision model is designed to maximize the robustness coefficient The goal is to take the bidding quantity and price of green electricity users as the decision-making target; the constraints include: the lower limit of the minimum profit of users' bids, the range of uncertainty variables of users' purchase of green certificates, the range of price uncertainty variables of users' participation in carbon quota trading, electricity purchase costs, electricity income, green certificate income, indirect carbon emission offset income, upper and lower limits of user quotations, and upper and lower limits of user reported quantities.

[0018] Preferably, the robustness coefficient is the green certificate price fluctuation range δ j,REC The fluctuation range of carbon quota price δ j,CO2 sum.

[0019] Preferably, the range of the uncertainty variable of the user purchasing the green certificate and the range of the price uncertainty variable of the user participating in the carbon quota transaction are respectively:

[0020]

[0021]

[0022] In the formula, is an uncertain set; π j,REC and π j,CO2 are the uncertainty variables of the actual prices for user j to purchase green certificates and participate in carbon quota trading; is the predicted value of user j for the green certificate price and carbon quota price during this green power trading cycle.

[0023] Preferably, the method for handling uncertainty is specifically as follows:

[0024] The adaptability of users to uncertainty fluctuations is measured through the robustness coefficient :

[0025]

[0026] In the formula: ω REC and ω REC are the weight coefficients of the fluctuation ranges of the green certificate price and carbon quota price respectively, and the sum of the two coefficients is 1.

[0027] Preferably, in step 2, the green power generator decision model aims to maximize the power generator's own power sales benefit, takes the green power generator's price and quantity reporting parameters as decision-making objectives, and takes the upper and lower limits of the green power generator's price offer and the upper and lower limits of the green power generator's quantity offer as constraints.

[0028] Preferably, in step 3, the green power trading clearing model clears with the goal of maximizing social welfare, takes the actual winning bid electricity quantities of green power generators and the actual winning bid electricity quantities of green power users as decision-making objectives, and the constraints are: power and electricity balance constraint, upper and lower limits constraint of the winning bid electricity quantity of green power generators, and upper and lower limits constraint of the winning bid electricity quantity of green power users.

[0029] Preferably, in step 4, the multi-MPEC model includes the MPEC model of green power generator i, the MPEC model of green power user j under deterministic conditions, and the MPEC model of green power user j considering uncertainty.

[0030] Preferably, in step 5, the improved diagonalization algorithm sets the initial value of the trading strategy, obtains the trading strategy of the green power generator by solving the MPEC model of green power generator i, obtains the optimal solution by solving the MPEC model of green power user j under deterministic conditions Based on The trading strategy of the green power user is obtained by solving the MPEC model of green power user j considering uncertainty, the trading strategy is updated, and the above process is iterated multiple times. The termination condition is that the update amounts of all quantity and price reporting information are less than the convergence threshold.

[0031] The second aspect of the present invention provides a system for constructing a green power bidding decision-making model considering prediction uncertainty using the method described in the first aspect of the present invention, including: a green power bidding decision-making model considering prediction uncertainty, a multi-MPEC model transformation module, and a multi-MPEC model solving module, characterized in that:

[0032] The module for constructing a green power bidding decision-making model considering prediction uncertainty: includes a green power user decision-making model considering uncertainty, a green power generator decision-making model, and a green power trading clearing model;

[0033] The multi-MPEC model transformation module: based on the primal-dual theorem, transforms the green power trading clearing model into the constraint conditions of the green power user decision-making model and the green power generator decision-making model to obtain a multi-MPEC model. The transformed model includes the MPEC model of green power generator i, the MPEC model of green power user j under deterministic conditions, and the MPEC model of green power user j considering uncertainty;

[0034] The module for solving the multi-MPEC model: linearizes the multi-MPEC model and uses an improved diagonalization algorithm to solve the MPEC model. Sets the initial value of the trading strategy, obtains the trading strategy of the green power generator by solving the MPEC model of green power generator i, and obtains the optimal solution by solving the MPEC model of green power user j under deterministic conditions Based on Obtains the trading strategy of the green power user by solving the MPEC model of green power user j considering uncertainty, updates the trading strategy and iterates the above process multiple times. The termination condition is that the update amounts of all reported volumes and quoted prices generated are less than the convergence threshold.

[0035] The beneficial effects of the present invention are as follows. Compared with the prior art,

[0036] (1) The present invention designs an optimization method and system for a user-side participation green power trading model affected by the carbon market, which has certain reference value for guiding how the user side participates in green power trading under the background of green power being deductible for indirect carbon emissions;

[0037] (2) The present invention designs a green power bidding decision-making model considering prediction uncertainty, which can objectively reflect the relative importance of uncertain variables in predicting green certificate and carbon quota prices, and effectively quantify the risk of prediction uncertainty;

[0038] (3) The present invention proposes an improved diagonalization algorithm, which simplifies the solution process of the market equilibrium problem of the robust optimization model based on information decision theory. Description of the Drawings

[0039] Figure 1It is the flowchart of the method and system for constructing a green power bidding decision model considering prediction uncertainty in the present invention;

[0040] Figure 2 It is the schematic diagram of the connection mode between green power trading and the carbon market in the present invention;

[0041] Figure 3 It is the flowchart of the improved diagonal algorithm in the present invention;

[0042] Figure 4 It is the comparison chart of the green power trading output results under Scenarios 1 and 2 of the embodiments of the present invention;

[0043] Figure 5 It is the sensitivity analysis chart of the revenue deviation factor on the L1 trading results in the embodiments of the present invention;

[0044] Figure 6 It is the sensitivity analysis chart of the regional power carbon emission factor on the trading results in the embodiments of the present invention. Detailed implementation manners

[0045] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. The embodiments described in this application are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the spirit of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0046] As shown in the Figure 1 accompanying drawings, the present invention proposes a method for constructing a green power bidding decision model considering prediction uncertainty, and the specific steps are as follows:

[0047] Step 1: Considering the uncertainty of the prediction of green certificate and carbon quota prices, construct a decision model for green power users; the benefits that users can obtain in a green power transaction mainly include the following three parts: electricity energy benefit, green certificate benefit, and indirect carbon emission offset benefit. The cost item is mainly the cost of purchasing green power to be paid.

[0048] 1) Power purchase cost

[0049] The power purchase cost R of user j j can be expressed as:

[0050] In the formula: is the clearing price of green power trading at time period t; T represents the total number of clearing time periods. Currently in China, generally 24 time periods are used to declare volume and price, that is, T = 24; is the winning bid electricity volume of green power user j at time period t; J is the set of all green power users.

[0051] 2) Electricity energy benefit

[0052] The electricity utility E of user j during period t j,t Can be approximately expressed by a quadratic function as follows:

[0053]

[0054] Where: and Are the coefficients of the corresponding terms.

[0055] The total electricity revenue obtained by user j in the transaction Can be expressed as

[0056]

[0057] 3) Green certificate revenue

[0058] Currently in China, green certificates are not allowed to be resold. The green certificate revenue obtained by users in green electricity trading is reflected in the cost of substituting the purchase of the same number of green certificates in the form of "separation of certificates and electricity" during the trading cycle. The total green certificate revenue obtained by user j in the transaction Can be expressed as:

[0059]

[0060] Where: Is the user The predicted value of the green certificate price during the current green electricity trading cycle.

[0061] 4) Indirect carbon emission offset revenue

[0062] User The deductible carbon emission volume A of the electricity purchased through this green electricity transaction j Can be calculated by the following formula:

[0063]

[0064] Where: Is the electricity carbon emission factor of region f.

[0065] Different from the fact that green certificates can only be traded once, carbon quotas allow repeated purchase and sale. For the carbon quota purchaser, the carbon offset revenue of the green electricity it purchases is equivalent to saving the cost of purchasing the corresponding amount of carbon quotas during this green electricity trading cycle. And as the seller, the carbon offset value of the green electricity purchased is equivalent to the additional profit obtained from selling the corresponding carbon quota increment in the carbon market during this cycle. The carbon offset revenue obtained by user j Can be expressed as:

[0066]

[0067] Where: Is the green electricity user The predicted value of the carbon quota price within the current green power trading cycle.

[0068] Without considering the errors in the predicted prices of green certificates and carbon quotas, user j bids in the green power trading with the goal of maximizing the power purchase benefit F j as follows:

[0069]

[0070] In the formula: represents the bid price of j at time t; and represent the upper and lower limits of the bid price of j at time t; represents the bid volume of j at time t; and represent the upper and lower limits of the bid volume of j at time t.

[0071] Not considering the uncertainty of the predicted prices of green certificates and carbon quotas will lead to a deviation between the predicted price value and the actual value, and this deviation will bring risks to the decision-making of the entity. Therefore, the present invention takes into account the risk tolerance of the user and balances the robustness and economy of its decision-making.

[0072] First, describe the uncertainty of user j of the predicted prices of green certificates and carbon quotas, and construct the corresponding uncertainty set:

[0073]

[0074] In the formula: is the uncertainty set; π j,REC , π j,CO2 are the uncertainty variables of the actual prices of user j for purchasing green certificates and participating in carbon quota trading; δ j,REC , δ j,CO2 are the fluctuation ranges of the prices of green certificates and carbon quotas.

[0075] By assigning different weights to δ j,REC , δ j,CO2 , the fluctuation ranges of the two can be normalized to the same coefficient. That is, by constructing the robustness coefficient to measure the adaptability of the user to the uncertainty fluctuation.

[0076]

[0077] ω REC +ω CO2 = 1 (13)

[0078] In the formula: ω REC , ω CO2 are the weight coefficients of the fluctuation ranges of the prices of green certificates and carbon quotas, respectively.

[0079] The differences and conflicts between indicators are reflected by calculating the standard deviation and correlation coefficient of the indicators. The specific steps are as follows:

[0080] 1) Select N groups of historical data sequence samples, and each group of samples contains two items of data: the average price of green certificates and the average price of carbon quotas in one cycle. First, normalize the m (m = 1, 2) - th data π in the n - th (n = 1, 2, …, N) sample mn to obtain the standardized matrix

[0081]

[0082] In the formula: is the data after normalization.

[0083] 2) Calculate the standard deviation to obtain the contrast σ between indicators m :

[0084]

[0085] 3) Calculate the correlation to obtain the conflict R between indicators m :

[0086]

[0087] 4) Calculate the information content S of the indicator m and the weight ω m :

[0088]

[0089] Establish a risk - averse user decision - making model with maximizing the robustness coefficient F j as the goal. The robust bidding decision - making model for green power users considering prediction uncertainty is shown in Equations (19) - (23):

[0090]

[0091] Equations (1) - (6), Equation (8), Equation (9), (23)

[0092] In the formula: m j is the revenue deviation factor of user j, which represents the deviation between the expected robust optimization goal and the optimal solution of the deterministic model; the larger its value, the greater the degree of risk aversion of the decision - maker; is the optimal solution of the deterministic green power user decision - making model.

[0093] Step 2: Construct a decision - making model for green power generators;

[0094] Green electricity requires a certain amount of operation and maintenance costs to maintain stable power supply. This invention assumes that the cost C of green electricity generator i at time period t i,t satisfies the following quadratic function relationship:

[0095]

[0096] In the formula: is the winning bid electricity quantity of generator i at time period t; and are the power generation cost coefficients; I is the set of all green electricity generators.

[0097] In green electricity trading, generator i aims to maximize its own electricity sales benefit:

[0098]

[0099] In the formula: F i is the total electricity sales benefit of generator i; represents the quotation of generator i at time period t; and represent the upper and lower limits of the quotation of generator i at time period t respectively; represents the quantity of electricity reported by generator i at time period t; and represent the upper and lower limits of the quantity of electricity reported by generator i at time period t respectively.

[0100] Step three: Construct a green electricity trading clearing model, specifically;

[0101] The trading center clears with the goal of maximizing social welfare, as shown in Equation (28). Equation (29) is the power and electricity balance constraint, and Equations (30) and (31) are the upper and lower limits constraints of the winning bid electricity quantities of green electricity generators and users respectively. Among them, the variable after the constraint condition is the dual variable corresponding to the corresponding constraint condition.

[0102]

[0103] In step four, the two-layer optimization model is transformed into a single-layer MPEC model, specifically;

[0104] Since the lower-layer clearing model is a convex problem, it can be transformed into the optimality conditions corresponding to the upper-layer model to realize the transformation from the two-layer optimization model to the single-layer MPEC model. The specific process is as follows:

[0105] 1) The MPEC model of green electricity user j under deterministic conditions:

[0106] min - F j (32)

[0107]

[0108] Among them, Equation (32) is the function for minimizing the negative power purchase benefit of green power user j; Equations (33) and (34) are the original constraints of the upper-layer model of the user; Equation (35) is the formula of the strong duality theorem; Equations (36)-(38) are the original constraints of the lower-layer trading model; Equations (39) and (40) are the optimality condition constraints of the lower-layer model; Equations (41) and (42) are the non-negativity constraints of the dual variables of the lower-layer model.

[0109] 2) MPEC model of green power user j considering uncertainty:

[0110]

[0111] s.t. Equations (20)-(23), Equations (33)-(42)(44)

[0112] Among them, Equation (43) is for user j to minimize its negative robustness coefficient.

[0113] 3) MPEC model of green power generator i:

[0114] min - F i (45)

[0115]

[0116] Among them, Equation (45) is the function for minimizing the negative power selling profit of the green power generator; Equations (46) and (47) are the original constraints of the upper-layer model of the generator; Equation (48) is the formula of the strong duality theorem; Equations (49)-(51) are the original constraints of the lower-layer trading model; Equations (52) and (53) are the optimality condition constraints of the lower-layer model; Equations (54) and (55) are the non-negativity constraints of the dual variables of the lower-layer model.

[0117] In Step 5, the MPEC model is linearized and the improved diagonalization algorithm is used to solve the MPEC model. Specifically:

[0118] The above MPEC model has non-linear terms. Non-linear terms caused by the multiplication of two decision variables exist in Equations (35) and (48), and a linearization method is required to decouple them. The steps are as follows:

[0119] 1) First, use the equivalence between the complementary slackness condition and the formula of the strong duality theorem for transformation:

[0120]

[0121]

[0122] 2) Use the big M method to decompose the complementary slackness condition in the form of 0 ≤ μ^P ≥ 0 into the following equivalent linear constraints:

[0123]

[0124] where: ν is a Boolean variable; M μ and M P are sufficiently large constants. Generally speaking, to solve the equilibrium problem composed of multiple MPECs, the generalized Lagrangian function corresponding to each decision maker can usually be constructed, and the KKT (Karush-Kuhn-Tucker) conditions of each MPEC are derived to obtain the feasible region of the equilibrium problem. And the profit of all decision makers or the true social welfare (TSW) of the electricity market is set to be maximized as the objective function of the EPEC (equilibrium problem with equilibrium constraints, EPEC), and then the market equilibrium solution that satisfies all the sets of optimality conditions is found. However, in the problem of this paper, since the user decision model is a robust optimization model, it is difficult to directly find a suitable objective function for the EPEC model. Therefore, the traditional method is not applicable to the solution of the equilibrium model proposed in this paper.

[0125] To address the above problems, this paper uses an improved diagonalization algorithm to solve the market equilibrium point. The optimal strategy is obtained by iteratively solving the MPEC models of each decision maker. The termination condition is that the differences between the quantity-price decisions of all entities in this round and those in the previous round are within the set acceptable range. At the same time, the maximum number of iterations is set. If convergence is not achieved after reaching the maximum number of iterations, "unable to converge to the equilibrium point" is output and the initial quantity-price of each strategy entity is adjusted and recalculated. When iterating to the MPEC model of green power users, the calculation steps are as follows:

[0126] 1) Without considering the uncertainties of the predicted green certificate price and carbon quota price in the decision-making of green power users, that is, δ j,REC and δ j,CO2 are both 0. Fix the bidding strategies of all other entities and solve equations (32)-(42) to obtain the optimal solution of the deterministic model

[0127] 2) Considering the uncertainties of the predicted green certificate price and carbon quota price, based on the revenue deviation factor of green power users, solve equations (43)-(44) to obtain the robust bidding strategy of this green power user.

[0128] Figure 2 Figure 26 is the flow chart of the algorithm for this paper, where n max 、i max 、j max are the maximum number of loops, the number of power generators participating in this green power transaction, and the number of users respectively.

[0129] The equilibrium of the electricity market is essentially the result of the mutual game and competition among market participants. In a healthy and competitive market, after multiple rounds of bidding by each participant, when all decision-makers are unwilling to update their bidding strategies, it is considered that the market has reached the Nash equilibrium. In one iteration, each green power generator and green power user will sequentially obtain corresponding strategies with the goal of maximizing their own benefits. After all participants have obtained new strategies in one iteration, a convergence judgment is made, that is, all the strategies obtained in this iteration are compared one by one with all the strategies of the previous iteration. If the convergence condition is met, it is considered that the Nash equilibrium state has been reached, and all market participants cannot obtain more benefits by updating their own bids. At this time, the decision-making update is completed and output. The above process is the corresponding process for each market participant to optimize their decisions. When a green power generator or user makes a decision, the quantity and price bidding information of all other market participants will not change and is regarded as known quantities, which are the latest quantity and price strategies obtained in the previous iteration.

[0130] First, determine which one or several strategy participants' quantity and price declaration strategies are difficult to converge, and compare the strategies generated in the last iteration with the initial strategies. Based on this, increase or reduce the initial quantity and price declaration strategies of each participant, and re-perform the iterative calculation. Considering the large computational volume of the above process, a parallel computing method can be used to accelerate the solution speed.

[0131] In the present invention, all the quantity declarations and price quotations generated in this iteration are set to have a difference of less than 0.01 from all the quantity declarations and price quotations generated in the previous round. That is considered the acceptable range. The maximum number of iterations of this patent is 2000 rounds.

[0132] Through a numerical example, verify the effectiveness of the proposed user-side participation in the green power trading bidding strategy in a multi-decision-maker game environment. It is assumed that 5 power generators (R1 to R5) and 5 users (L1 to L5) participate in the monthly centralized competitive bidding for green power. The predicted value of the monthly green certificate purchase price for users is uniformly set at 45 yuan / MWh; the predicted value of the monthly carbon quota trading price is uniformly set at 82 yuan / t. Through calculation, the weights of the uncertain variables are 0.39 and 0.61 respectively. The regional electricity carbon emission factor is taken as 0.87 tCO2 / MWh. The detailed parameters of market participants are shown in Appendix A.

[0133] To explore the impact of the carbon market on the clearing results of green power trading under the background of deductible indirect carbon emissions, the following scenarios are set:

[0134] 1) Scenario 1, there is no connection mechanism between green power trading and the carbon market, that is, the indirect carbon emissions are also counted when users purchase green power;

[0135] 2) Scenario 2, there is a connection between green power trading and the carbon market, and the regional average electricity emission factor is used to calculate and deduct the indirect carbon emissions.

[0136] Under the market equilibrium state, the winning electricity quantities and clearing electricity prices of green electricity users in each period under the two scenarios are as Figure 3 shown.

[0137] It can be Figure 3 seen that compared with Scenario 1, the clearing electricity price in Scenario 2 increases and the winning electricity quantities of each user increase to varying degrees. Select the fifth period for specific analysis. The clearing electricity price in Scenario 1 is 284.76 yuan / MWh, and the clearing electricity price in Scenario 2 is 319.17 yuan / MWh, which is a 12.08% increase compared with Scenario 1. The total cleared electricity quantity increases by 362.61 MWh. The growth of the winning electricity quantities of each green electricity user in this period is shown in Table 1. The benefit deviation factors of users L4 and L5 are relatively small, indicating that their risk resistance capabilities are relatively strong. They will increase their quantity and price declarations to a greater extent in this green electricity transaction to obtain more market share and pursue high-carbon market benefits. However, the benefit deviation factors of L1, L2, and L3 are relatively large, indicating that these three users have a greater degree of risk aversion. Therefore, they will make relatively conservative decisions, with a smaller increase in their bids, and the growth of their winning electricity quantities is significantly less than that of L4 and L5, that is, sacrificing their acceptable expected benefits in exchange for strong robustness of their bidding strategies. The above analysis shows that the background of deductible indirect carbon emissions is conducive to stimulating the potential of green electricity consumption on the demand side. As more users purchase green electricity to pursue carbon offset value, it will promote a decrease in the carbon emissions of regional electricity consumption. The degree of risk aversion of each user to the uncertainty of the predicted price will affect the impact of the carbon market on their bidding strategies and clearing results.

[0138] Table 1 The bidding quantity of each green electricity user in the fifth period

[0139] Tab.1The bidding quantity of each green electricity user in the fifth period

[0140] Unit: MWh

[0141]

[0142] In Scenarios 1 and 2, the electricity sales revenues on the power generation side are shown in Table 2. It can be seen that the electricity sales revenue of power generators in Scenario 2 is higher than that in Scenario 3, indicating that the environment of deductible indirect carbon emissions of green electricity can also improve the enthusiasm of the supply side to participate in green electricity transactions, which is conducive to expanding the scale of green electricity transactions on the source side.

[0143] Table 2 Revenue comparison of green power generators under Scenarios 1 and 2

[0144] Tab.2Revenue comparison of green power generators under Scenarios 1 and 2

[0145] Unit: 10,000 yuan

[0146]

[0147]

[0148] The robustness analysis of the bidding strategy for users to participate in green power trading is as follows:

[0149] Taking L1 as an example, keeping the settings of other parameters unchanged, under different minimum expected revenue targets, the robustness coefficient and the expected revenue vary with the deviation factor m1 as Figure 4 shown. It can be seen that within a certain range, the robustness coefficient of predicting the prices of green certificates and carbon quotas increases with the increase of the revenue deviation factor. When the lower the expected revenue that green power users can accept, the better the robustness of their bidding strategy, and the more they can withstand the price fluctuations of green certificates and carbon quotas in a larger range.

[0150] When m1 = 0, the robustness coefficient is 0. This situation corresponds to L1 considering that the predicted value of the uncertain quantity is exactly the same as the actual value. As m1 increases, the robustness coefficient starts to change from 0, which indicates that when green power users bid at this time, compared with revenues with strong uncertainties such as green certificate and carbon offset revenues, they pay more attention to considering the value of electric energy and the cost of purchasing electricity, at the cost of reducing the expected total revenue and the total amount of purchased green power, to obtain a stronger robustness of the bidding strategy to cope with the deviation of the uncertain quantity.

[0151] When m1 = 0.3, in this case, that is, when the actual purchase price of green certificates is 19.11% lower than the predicted value and the actual carbon quota trading price is 29.89% lower than the predicted value, the revenue from purchasing green power obtained by L1's decision can ensure not less than 70% of the revenue of the deterministic model, that is, not less than 614,500 yuan.

[0152] When m1 increases to 0.6, the robustness coefficient reaches the peak value of 1. At this time, when the revenue deviation factor continues to increase, the robustness coefficient basically remains unchanged, and the constraint conditions related to the uncertain quantity will no longer be the effective constraint boundaries. At this time, even if L1 continues to reduce the expected total revenue, it cannot obtain a more robust bidding decision. Therefore, in this green power trading, L1 should not set m1 to a value greater than 0.6 to avoid being unable to flexibly balance the economy and robustness of its bidding strategy. The analysis of other users is similar and will not be elaborated here.

[0153] The sensitivity analysis of the regional power carbon emission factor to the clearing result is as follows:

[0154] The indirectly reduced carbon emissions that can be offset by the green electricity purchased by the user side are directly related to the regional electricity carbon emission factor. Different values will directly affect the user's bidding strategy in the green electricity trading, and then affect the clearing result. To explore its impact, the regional electricity carbon emission factor is set to increase from 0.65 to 0.95 with a step size of 0.05, and the model is solved on the premise that other parameters remain unchanged. The clearing volume and price in the 9th period and the trading results of L2 and L5 are selected for specific analysis, as Figure 5 shown.

[0155] From the perspective of the global clearing result, as the regional electricity carbon emission factor increases, both the clearing electricity price and the winning bid quantity show an upward trend. This is because the electricity carbon emission factor is positively correlated with the carbon offset quantity per unit of green electricity. Each user will change its bidding strategy to obtain more carbon offset benefits, thus changing the clearing result. It shows that in the context where green electricity is not included in the indirect carbon emission accounting, the bidding decisions of green electricity users in regions with higher electricity carbon emission levels change more significantly than those in lower regions, and the increase in the clearing volume and price is more obvious; users in high-carbon emission level regions will be more inclined to purchase green electricity, which will reduce the output space of emission units such as thermal power. In order to pursue higher benefits in the electricity market, they will be motivated to reduce their own carbon emission levels, which helps to promote the decline of the overall regional electricity carbon emission level. For individuals, L5 has a higher ability to bear the risk of predicting the carbon quota price than L2. Therefore, the change in the regional electricity carbon emission factor has a greater impact on L5 than on L2, indicating that the lower the decision-maker's degree of risk aversion, the more sensitive they will be to changes in the emission factor. In summary, the value of the regional electricity carbon emission factor has a great impact on the bidding strategy and clearing result of the user side participating in green electricity trading. Therefore, relevant departments should consider shortening the accounting cycle of the electricity carbon emission factor, which can not only truly reflect the emission reduction value of green electricity, guide the user side to make reasonable decisions in green electricity trading, but also help to ensure the fair and stable development of the market. The following conclusions can be drawn through the case study analysis:

[0156] 1) Under the influence of the carbon market, not including the green electricity trading volume in the indirect carbon emissions helps to guide the low-carbon transformation of enterprises' electricity consumption, improve the enthusiasm of both bilateral parties to participate in green electricity trading, and further expand the scale of green electricity trading.

[0157] 2) As a key link connecting the carbon market and green electricity trading, the value of the regional electricity carbon emission factor will have a great impact on the bidding strategy and clearing result of green electricity users. Relevant departments should consider appropriately shortening the accounting cycle to correctly reflect the carbon emission heterogeneity between green electricity and other electricity.

[0158] The beneficial effects of the present invention are that, compared with the prior art,

[0159] (1) The present invention designs an optimization method and system for the user-side participation in green power trading model affected by the carbon market, which has certain reference value for guiding how the user-side participates in green power trading under the background of green power deductible indirect carbon emissions;

[0160] (2) The present invention designs a green power bidding decision-making model considering prediction uncertainty, which can objectively reflect the relative importance degree of the uncertain variables of green certificate and carbon quota price prediction, and effectively quantify the risk of prediction uncertainty;

[0161] (3) The present invention proposes an improved diagonalization algorithm, which simplifies the solution process of the market equilibrium problem of the robust optimization model based on information decision theory.

[0162] The present disclosure may be a system, a method, and / or a computer program product. The computer program product may include a computer-readable storage medium having thereon computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0163] The computer-readable storage medium may be a tangible device that can retain and store instructions for use by an instruction execution device. The computer-readable storage medium may be, for example, but is not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disc (DVD), a memory stick, a floppy disk, a mechanically encoded device such as a punch card or raised structures in grooves having instructions stored thereon, and any suitable combination of the foregoing. The computer-readable storage medium as used herein is not construed as being a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagated through a waveguide or other transmission medium (e.g., an optical pulse through an optical fiber cable), or an electrical signal transmitted through a wire.

[0164] The computer-readable program instructions described herein may be downloaded from the computer-readable storage medium to various computing / processing devices, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network may include a copper transmission cable, an optical fiber transmission, a wireless transmission, a router, a firewall, a switch, a gateway computer, and / or an edge server. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing / processing device.

[0165] The computer program instructions for performing the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine - related instructions, microcode, firmware instructions, state - setting data, or source code or object code written in any combination of one or more programming languages, including object - oriented programming languages such as Smalltalk, C++, etc., and conventional procedural programming languages such as the "C" language or similar programming languages. The computer - readable program instructions may be executed entirely on the user's computer, partially on the user's computer, executed as a stand - alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or it may be connected to an external computer (e.g., through the Internet using an Internet service provider). In some embodiments, by using the status information of the computer - readable program instructions to customize an electronic circuit, such as a programmable logic circuit, a field - programmable gate array (FPGA), or a programmable logic array (PLA), the electronic circuit can execute the computer - readable program instructions to implement various aspects of the present disclosure.

[0166] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications or equivalent substitutions can still be made to the specific embodiments of the present invention, and any modification or equivalent substitution that does not depart from the spirit and scope of the present invention should be covered by the protection scope of the claims of the present invention.

[0167] Appendix A:

[0168] Table A1 User revenue deviation factor

[0169] Tab.A1 Users revenue deviation factor

[0170]

[0171] Table A2 Parameters of users(1)

[0172] Tab.A2 Parameters of users(1)

[0173]

[0174] Table A3 Parameters of users(2)

[0175] Tab.A3 Parameters of users(2)

[0176]

[0177]

[0178] Table A4 Parameters of generators(1)

[0179] Tab.A4 Parameters of generators(1)

[0180]

[0181] Table A5 Parameters of generators(2)

[0182] Tab.A5 Parameters of generators(2)

[0183]

[0184]

Claims

1. A method for constructing a green electricity bidding decision-making model considering prediction uncertainty, characterized in that, Including: Step 1: Considering the uncertainty of green certificate and carbon quota price prediction, construct a decision-making model for green electricity users; Step 2: Construct a decision-making model for green electricity generators; Step 3: Construct a clearing model for green electricity trading; Step 4: Based on the primal-dual theorem, transform the constructed models into a multi-MPEC model; Step 5: Linearize the multi-MPEC model and use an improved diagonalization algorithm to solve the MPEC model.

2. The method for constructing a green electricity bidding decision-making model considering prediction uncertainty according to claim 1, characterized in that: In Step 1, the green power user decision-making model aims to maximize the robustness coefficient and takes the bid volume and price of green power users as the decision-making objectives. The constraints include: the lower limit constraint of the minimum revenue of user bids, the range of the uncertainty variables of the green certificates purchased by users, the range of the uncertainty variables of the carbon quota trading price participated by users, the power purchase cost, the electricity revenue, the green certificate revenue, the revenue from offsetting indirect carbon emissions, the upper and lower limits of user quotes, and the upper and lower limits of user reported quantities.

3. The method for constructing a green electricity bidding decision-making model considering prediction uncertainty according to claim 2, characterized in that: The robustness coefficient is the sum of the green certificate price fluctuation range δ j,REC and the carbon quota price fluctuation range δ j,CO2 .

4. The method for constructing a green electricity bidding decision-making model considering prediction uncertainty according to claim 2, characterized in that: The ranges of the uncertainty variables for the user's purchase of green certificates and the uncertainty variables for the user's participation in the carbon quota trading price are respectively: In the formula, is an uncertain set; π j,REC , π j,CO2 are the uncertainty variables of the actual prices for user j to purchase green certificates and participate in carbon quota trading; is the predicted value of user j for the green certificate price and carbon quota price during this green power trading cycle.

5. The method for constructing a green electricity bidding decision-making model considering prediction uncertainty according to claim 4, characterized in that: The method for dealing with uncertainty, specifically: Through the robustness coefficient Measure the adaptability of users to uncertainty fluctuations: Where: ω REC and ω REC are the weight coefficients of the price fluctuation ranges of green certificates and carbon quotas respectively, and the sum of the two coefficients is 1.

6. The method for constructing a green electricity bidding decision-making model considering prediction uncertainty according to claim 1, characterized in that: In step 2, the decision-making model of the green electricity generator aims to maximize the power seller's own power sales benefit, takes the bid price and bid volume of the green electricity generator as the decision-making objectives, and takes the upper and lower limits of the generator's bid price and the upper and lower limits of the generator's bid volume as the constraint conditions.

7. The method for constructing a green electricity bidding decision-making model considering prediction uncertainty according to claim 1, characterized in that: In step 3, the clearing model for green electricity trading has the following objective function: Taking the actual winning bid electricity volume of the green electricity generator and the actual winning bid electricity volume of the green electricity user as the decision-making objectives, the constraint conditions are: power and electricity balance constraint, upper and lower limits constraint of the winning bid electricity volume of the green electricity generator, and upper and lower limits constraint of the winning bid electricity volume of the green electricity user.

8. The method for constructing a green electricity bidding decision-making model considering prediction uncertainty according to claim 1, characterized in that: In step 4, the multi-MPEC model includes the MPEC model of green electricity generator i, the MPEC model of green electricity user j under deterministic conditions, and the MPEC model of green electricity user j considering uncertainty.

9. The method for constructing a green electricity bidding decision-making model considering prediction uncertainty according to claim 1, characterized in that: In step 5, for the improved diagonalization algorithm, set the initial value of the trading strategy, obtain the trading strategy of the green power generator by solving the MPEC model of green power generator i, and obtain the optimal solution by solving the MPEC model of green power user j under deterministic conditions. Based on Solve the MPEC model of green power user j considering uncertainty to obtain the trading strategy of the green power user, update the trading strategy and iterate the above process multiple times. The termination condition is that the update amounts of all quantity declarations and price quotes are less than the convergence threshold.

10. A green electricity bidding decision model construction system that uses the method according to any one of claims 1-9 and considers prediction uncertainty, comprising: Construct a green electricity bidding decision-making model considering prediction uncertainty, a multi-MPEC model transformation module, and a module for solving the multi-MPEC model, characterized in that: The module for constructing a green electricity bidding decision-making model considering prediction uncertainty: includes a decision-making model for green electricity users considering uncertainty, a decision-making model for green electricity generators, and a clearing model for green electricity trading; The multi-MPEC model transformation module: Based on the primal-dual theorem, transform the clearing model for green electricity trading into the constraint conditions of the decision-making model for green electricity users and the decision-making model for green electricity generators to obtain a multi-MPEC model. The transformed model includes the MPEC model of green electricity generator i, the MPEC model of green electricity user j under deterministic conditions, and the MPEC model of green electricity user j considering uncertainty; Multi-MPEC model solving module: Linearize the multi-MPEC model, solve the MPEC model using the improved diagonalization algorithm, set the initial value of the trading strategy, obtain the trading strategy of the green power generator by solving the MPEC model of green power generator i, and obtain the optimal solution by solving the MPEC model of green power user j under deterministic conditions. Based on Solve the MPEC model of green power user j considering uncertainty to obtain the trading strategy of the green power user, update the trading strategy and iterate the above process multiple times. The termination condition is that the update amounts of all reported volumes and quoted prices generated are less than the convergence threshold.

Images