A method and system for power generator carbon market decision-making that considers cash flow

Abstract

This invention discloses a decision-making method and system for power generators in the carbon market that considers cash flow, relating to the field of decision optimization in the power market and carbon trading market for power companies. It addresses the collaborative decision-making problem and cash flow modeling problem in the power market and carbon market for power companies. The method includes the following steps: establishing a year-round stochastic decision-making model for the power-carbon market with cash flow constraints; then, using mixed time-step modeling to reduce decision variables while still describing the cash settlement cycle, and simplifying the multi-stage stochastic optimization model into a two-stage stochastic optimization model for easier solution; finally, addressing the problem of frequent speculative trading in the second stage of the traditional two-stage model by adding a one-way carbon market trading rule, and addressing the problem of the two-stage model's underestimation of future carbon purchase costs by adding a timing trading rule. This method fully utilizes the characteristics of the power-carbon market decision-making process, ensures cash flow security, effectively reduces the carbon purchase costs for power generators, and improves overall profitability.

Description

Technical Field

[0001] This invention relates to the field of decision optimization for power companies in the electricity market and carbon trading market, and in particular to a method and system for power generation companies to make decisions in the carbon market that takes cash flow into account. Background Technology

[0002] In the context of the electricity-carbon market, generating units are no longer merely physical resources subject to unified grid dispatch. The primary intentions and operating conditions of power generators become indispensable factors. It is necessary to re-examine the electricity-carbon market decision-making of power generators from the perspective of the Cyber-Physical-Social System in Energy (CPSSE). All activities of power generators generate cash flow; insufficient cash not only leads to late payments and penalties but can also disrupt normal operations, posing significant risks to the enterprise. Figure 1 As shown, power generators' decisions in the electricity market and carbon trading market are coupled. The annual settlement of carbon allowances necessitates considering the entire year's situation when making spot market decisions. Since power generators can purchase all or part of their required annual carbon allowances on any trading day before settlement, and can also sell their existing allowances, and given the constantly changing and uncertain nature of carbon allowance prices, the final carbon purchase cost for a power generator in a year will depend on its carbon market decision-making strategy. The carbon price on the day of the power generation decision does not directly represent the actual carbon emission cost. This provides greater flexibility to power generators' daily electricity-carbon spot market decisions but also increases complexity. Existing research, such as the literature "JIANG Kai, LIU Nian, YAN Xiaohe, et al. Modeling Strategic Behaviors for GenCo with joint Consideration on Electricity and Carbon Markets," does not consider the annual settlement of carbon allowances, using the daily carbon price or long-term average carbon price to calculate emission costs. This simplifies the problem by requiring daily balance between emissions and allowances from a single-day perspective, but it also does not consider the cash flow constraints of power generators. Such methods cannot fully utilize the decision-making flexibility of the electricity-carbon market to achieve better decisions, and may also result in decisions that fail to meet cash flow constraints, making the decisions unenforceable. Summary of the Invention

[0003] To address the issues of coordinated decision-making and cash flow modeling in the electricity and carbon markets for power companies, this invention discloses a method and system for optimizing the decision-making process of power generation companies in the electricity and carbon spot markets, taking into account cash flow constraints and carbon emission quota settlement cycles.

[0004] To achieve the objectives of this invention, the present invention provides a power generator carbon market decision-making method that considers cash flow, comprising the following steps:

[0005] Step 1: Considering the diverse cash settlement cycles of operating activities, establish a stochastic decision-making model for the electricity-carbon market with a full-year perspective and cash flow constraints;

[0006] Step 2: Use mixed time step modeling to reduce the number of decision periods and simplify the annual view of the stochastic decision-making model of the electricity-carbon market into a two-stage stochastic optimization model;

[0007] Step 3: Add carbon market one-way trading rules and carbon market timing trading rules to the annual perspective electricity-carbon market stochastic decision-making model. The carbon market one-way trading rule assumes that generators have a risk-averse tendency in the carbon market, do not speculate in the carbon trading market, and only conduct one-way trading based on future expectations. The carbon market timing trading rule is to buy and sell carbon allowances based on market carbon price, purchase price and selling price.

[0008] Step 4: Solve the problem to obtain the decisions for the electricity market and carbon trading market.

[0009] The present invention also provides a power generator carbon market decision-making system that considers cash flow, for implementing the aforementioned method, the system comprising the following modules:

[0010] The decision model building module is used to build a stochastic decision model for the electricity-carbon market with a full-year perspective and cash flow constraints.

[0011] A simplification module is used to reduce the number of decision periods by adopting mixed time step modeling and to simplify the annual view stochastic decision-making model of the electricity-carbon market into a two-stage stochastic optimization model.

[0012] The rule-building module is used to add carbon market one-way trading rules and carbon market timing trading rules to the annual view electricity-carbon market stochastic decision-making model. The carbon market one-way trading rule assumes that generators have a risk-averse tendency in the carbon market, do not speculate in the carbon trading market, and only conduct one-way trading based on future expectations. The carbon market timing trading rule is to buy and sell carbon allowances based on market carbon price, purchase price and selling price.

[0013] The solver module is used to perform the solution and obtain the decisions for the electricity market and carbon trading market.

[0014] Compared with the prior art, the beneficial effects of the present invention are:

[0015] The method of this invention makes full use of the decision-making characteristics of the electricity carbon market, which can effectively reduce the carbon purchase cost of power generators and improve overall revenue. At the same time, it takes into account cash flow constraints and can avoid decisions that cannot be executed due to insufficient cash. Attached Figure Description

[0016] Figure 1 This is a schematic diagram illustrating the coupling between electricity and carbon market decisions.

[0017] Figure 2 This is a diagram illustrating the settlement cycle for different activities.

[0018] Figure 3 This is a daily rolling decision-making flowchart for the electricity-carbon market for power generators throughout the year. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] This invention first considers the diverse cash settlement cycles of business activities and establishes a year-round stochastic decision-making model for the electricity-carbon market with cash flow constraints. Then, it adopts mixed time step modeling to reduce decision variables while still being able to describe the cash settlement cycle, and simplifies the multi-stage stochastic optimization model into a two-stage stochastic optimization model for easier solution. Finally, it addresses the problem of frequent speculative trading in the second stage of the traditional two-stage model by adding one-way carbon market trading rules, and addresses the problem of the two-stage model having an underestimation of future carbon purchase costs by adding carbon market timing trading rules.

[0021] Step 1: Establish a stochastic decision-making model for the electricity-carbon market with a full-year perspective and cash flow constraints.

[0022] The steps for establishing a stochastic decision-making model for the electricity-carbon market with a full-year perspective and cash flow constraints include:

[0023] The goal of power generators in making daily market decisions is to maximize the total annual revenue converted to that day. To achieve this, they need to consider both the electricity-carbon market decisions for the day and the future market trading plans. Without considering the settlement cycle, the daily optimization objective function of power generators is shown in equation (1).

[0024]

[0025] d∈{d0,2,...,D},D∈{365,366} (2)

[0026]

[0027]

[0028] Where d is the index of each decision day within the year, d0 is the index of the current decision day (rolling from 1 to D), and D is the index of the end of the year (365 for common years and 366 for leap years); The revenue obtained up to day d0 is a known amount; Electricity market sales revenue, generation costs, carbon market carbon purchase expenditures, loan repayments, and daily expenses for day d, respectively. dis <1 represents the daily discount factor, c Penalty Fines incurred for unpaid emission quotas; d It is a vector of random variables over day d. The electricity price at time h on day d. Let δ be the carbon price on day d. d For daily expenses payable on day d; x d It is the vector of decision variables on day d, p d,h,g The output of generator set g at time h on day d. This represents the carbon allowance trading volume on day d (positive for purchases). This represents the number of loan applications per day (d). This represents the loan repayment amount on day d. This refers to the outstanding late payment due on day d.

[0029] Since the various business activities of generators have different settlement cycles, and their optimization problem is a constrained problem, the discount factor is not a constant and is difficult to express explicitly. After considering the settlement cycle, this invention sets the daily optimization decision objective as maximizing the expected total assets of the generator at the end of the year, as shown in Equation (5). The time value of cash is implied in the cash flow constraint below.

[0030]

[0031] in, It is the cash you have at the end of the year, R D The electricity bill had not yet arrived by the end of the year. and It refers to outstanding loans at the end of the year and their penalty coefficients. and It represents the remaining amount of delayed payment at the end of the year and its penalty coefficient.

[0032] Power generators make profits by selling electricity in the electricity market. While generating electricity, they consume fuel and must also comply with the operating constraints of the generator units.

[0033]

[0034]

[0035]

[0036]

[0037] in, The electricity cost for the generator on day d is given, where Δh represents the duration of one hour, used to unify the dimensions of the left and right sides of the equation. The electricity price at time h on day d, p d,h,g The generating capacity of unit g at h on day d. The cost of fuel consumed by the power generator for generating electricity on a given day. denoted as the cost coefficient for unit g. and p g These are the upper and lower power limits for unit G, respectively. and These represent the maximum hourly uphill and downhill gradients for generator unit g. The generator units here can correspond to physical units or be an equivalent model combining multiple units.

[0038] Greenhouse gas emissions are generated during power generation. Under the carbon trading mechanism, power generators need to purchase allowances for their emissions and settle the full amount of allowances in the following year, otherwise they will face penalties. This invention does not consider the issue of multi-year allowance coordination, therefore the settlement date is set as the end of the year, D.

[0039]

[0040]

[0041]

[0042]

[0043]

[0044] in The net carbon emissions of generators per day after deducting the free rate. and These represent the emissions generated per unit of electricity generated by unit g and the free allowance determined by the baseline method, respectively. This refers to the total emission allowances held by the generators on day d. This represents the amount of carbon market allowances traded by the generator on day d (positive for buying, negative for selling). This represents the total net emissions generated by power generators from the beginning of the year to date. This is the net emissions generated by the daily power generation company. The fees incurred for the trading quota on day d. The price of carbon allowances on day d. Penalty For the penalties incurred due to unpaid emission quotas, λ Penalty The penalty for the unit's unpaid quota. This represents the total annual carbon emissions. Carbon allowances held by power generators at the end of the year.

[0045] In the early stages of carbon trading, the proportion of free allowances was relatively high, resulting in a smaller increase in power generation costs for generators due to carbon emissions. As the mechanism matures and emission limits tighten, my country will gradually reduce the proportion of free allowances. The European Emissions Trading System (EU ETS) eliminated free allowances for the power generation sector in 2013 due to instances where free allowances generated unexpected profits for generators (windfall profits). Based on these trends, in some embodiments of the present invention...

[0046] Power generators require financial support for all their activities. When cash runs out, some activities cannot be carried out, leading to losses. Therefore, a cash flow model for power generators is established, as shown in the attached figure. Figure 2 As shown.

[0047]

[0048]

[0049]

[0050]

[0051]

[0052]

[0053]

[0054]

[0055]

[0056] in For the cash reserves on day d, u Cash The daily interest rate for deposits. This is the set of power generation dates for which electricity bills are paid on day d. In some embodiments of the present invention, the electricity bill for the previous month is set to be paid on the 10th of each month; therefore, when d is the 10th of a certain month... Including all decision dates of the previous month, when d is not the 10th of a certain month. It is an empty set. For the set of power generation dates for which fuel costs are paid on day d, in some embodiments of the invention, fuel costs for the period from the 1st to the 15th of the current month are paid on the 20th of each month, and fuel costs for the period from the 16th to the end of the previous month are paid on the 5th of each month. To advance Δd LA Loans applied for on day d and disbursed on day d, of which Δd LA The time required to approve a loan application is set to 7 days in some embodiments of the present invention. The loan to be repaid on day d, For the overdue payment due on day d, δ d This refers to the daily expenses for day d. For the cumulative delayed payment amount over day d, u Delay This is the daily late payment fee rate. The total outstanding loan amount on day d, u Loan The daily interest rate for the loan. This is the maximum loan amount for the company. Δd is the minimum repayment amount for a loan on day d. LMR The longest period for a single loan is 30 days in some embodiments of the present invention. Δd is the minimum repayment amount for a loan on day d. LCR The shortest loan period is 7 days in some embodiments of the present invention. The electricity cost corresponding to the power generation on day d′ Here, d′ represents the fuel cost for power generation on day d′, and d′ is the index of the transaction day being considered. For loans applied for on date d-Δd′, Δd′ is The corresponding number of days since the loan application dated d. For d-Δd LMR -Δd LA Loans applied for on the same day For d-Δd LCR -Δd LA Loan applied for on the same day.

[0057] Equation (15) represents the daily change in cash reserves. Equation (16) indicates that cash reserves cannot be negative. Equation (17) represents the daily change in the total amount of deferred payments. Equation (18) indicates that only fuel costs, loan repayments, and daily expenses can be deferred; activities such as purchasing carbon credits are not permitted when cash is unavailable. Equation (19) represents the daily change in the total amount of outstanding loans. Equation (20) indicates that the sum of the current total loan amount and the outstanding loans applied for but not yet disbursed is non-negative and less than the maximum loan amount. Equations (22), (23), and (21) represent the impact of loan repayment cycle restrictions on the upper and lower limits of daily repayments.

[0058] Step 2: Use mixed time step modeling to reduce the number of decision periods, and simplify the annual view stochastic decision-making model of the electricity-carbon market established in Step 1 into a two-stage stochastic optimization model.

[0059] First, a hybrid time step modeling approach is adopted to reduce the number of decision-making periods. The idea is to expand the decisions of the most recent two months to daily to take into account the impact of the settlement cycle, and aggregate subsequent periods to monthly based on typical days to reduce the problem size.

[0060] Even after adopting mixed time step modeling, the number of decision stages is still relatively large. For example, the decision problem on January 1st has 31+28+10=69 stages, making it difficult to directly build a scenario tree. Therefore, the multi-stage stochastic optimization model is simplified to a two-stage stochastic optimization model, that is, only distinguishing between the current (Here and Now) and future (Wait and See) stages, as shown in equation (24):

[0061]

[0062] in Let represent the decision variable and random variable from day d0 to day D, respectively, and f be the objective function. Further processing using the scenario method yields equation (25), where N is the total number of scenarios and n is the scenario index.

[0063]

[0064] Let be the decision variables from day d0+1 to day D in scenario n. Let be the decision variables from day d0 to day D in scenario n. Let be the value of the random variable from day d0 to day D in scenario n.

[0065] However, directly applying two-stage stochastic optimization models to carbon trading decision-making has limitations. Compared to multi-stage models, two-stage models ignore the bifurcation scenario on day 2 and beyond, utilizing complete future information in subsequent decision-making stages, thus losing nonanticipativity. Furthermore, carbon trading volumes can be transferred over a considerable period before liquidation, leading to two problems in carbon trading decision-making:

[0066] (1) It leads to unrealistic and frequent speculative trading. Because carbon prices fluctuate daily, it is assumed that if the complete future carbon price curve is known, one can frequently buy low and sell high to make a profit, but such perfect price prediction does not exist in reality.

[0067] (2) Underestimation of the future cost of carbon allowance purchases. Since there are no unforeseen factors, the second phase will make carbon allowance purchase decisions based on the complete future carbon price curve, purchasing the required carbon allowances when the predicted carbon price is at its lowest, leading to decision-makers underestimating the future cost of carbon allowance purchases.

[0068] The above problem is caused by the overfitting of the decision-making strategy to a limited sample scenario. Although it can achieve good results in the sample scenario, the effect is not ideal in actual rolling optimization.

[0069] Step 3: Add one-way trading rules to the carbon market to address the problem of frequent speculative trading in the second stage of the traditional two-stage model, and add timing trading rules to the carbon market to address the problem of the two-stage model having low expectations for future carbon purchase costs.

[0070] To avoid overfitting of carbon trading decision-making strategies to sample scenarios, this invention narrows the strategy space by adding constraint rules to the decision-making model, including carbon market one-way trading rules and carbon market timing trading rules.

[0071] First, a one-way carbon trading rule is added to eliminate active speculation. This assumes that power generators have a risk-averse tendency in the carbon market and do not speculate, only engaging in one-way trading based on future expectations (buying only scarce allowances or selling surplus allowances in the same scenario, without buying first and then selling or selling first and then buying). The specific form of the one-way carbon trading rule is shown in equations (26-18). Since power generation emissions and carbon trading volume are coupled, changes in the direction of carbon trading may alter the marginal cost of power generation emissions. Therefore, the trading direction cannot be simply determined based on the current allowance holdings, but rather optimized using integer programming methods.

[0072]

[0073]

[0074]

[0075] in, It is a 0-1 integer variable. A value of 1 indicates that the carbon quota trading direction in the nth scenario is buying, otherwise it is selling. M represents the carbon allowance trading volume for day d under scenario n. Sell and M Buy These are constants, representing the maximum possible future quota selling and buying volumes, respectively. In this invention, they are taken as the quota trading volumes corresponding to the power generator's decision-making periods of no power generation and full power generation.

[0076] Equation (27) indicates that when the trading direction of scenario n is buying, the sum of the absolute values ​​of daily trading volumes equals the sum of daily trading volumes; otherwise, the constraint is relaxed. Equation (28) indicates that when the trading direction of scenario n is selling, the sum of the absolute values ​​of daily trading volumes equals the negative of the sum of daily trading volumes; otherwise, the constraint is relaxed. The reason for constraining the sum of daily trading volumes rather than the single-day trading volume is that this helps to reduce the relaxation degree of the Big M method, thereby accelerating the solution of integer programming problems by the branch and bound method.

[0077] Secondly, the problem of underestimated future carbon purchase costs is addressed by adding carbon market timing trading rules, thus preventing power generators from continuously postponing their carbon purchases. The main idea is as follows: first, a parameterized timing trading rule is given; then, a heuristic optimization algorithm is used to optimize the corresponding parameters in numerous multi-stage carbon trading scenario simulations; finally, the obtained timing trading rule is added to the two-stage model of this invention. The form of the carbon market timing trading rule is as follows:

[0078] (1) The market carbon price is less than or equal to the purchase price. Carbon allowances can be purchased at any time;

[0079] (2) The market carbon price is greater than or equal to the selling price. Carbon allowances can be sold at any time;

[0080] (3) Market carbon price at the purchase price and selling price Carbon quota trading will not be conducted during this period;

[0081] (4) Carbon allowances can be freely bought and sold on the last trading day.

[0082] Among them, the carbon trading purchase price on day d and selling price As shown in equations (29) and (30), α Buy β Buy α Sell β Sell These are parameters to be determined.

[0083]

[0084]

[0085] Purchase price and selling price The meanings are as follows: when considering the correlation between the final transaction price and the transaction strategy, the expected transaction price for buying and selling a unit of carbon quota according to the given carbon trading strategy during the tradable period from day d+1 to the end of the year. The above rule determines whether to trade on day d by comparing the carbon price on day d with the expected future transaction price. Equations (29) and (30) are linear fits of the expected future transaction price with respect to d (the shorter the remaining tradable time, the higher the acceptable buying price and the lower the selling price). This invention combines the Monte Carlo method and the differential evolution algorithm to solve the following problems to obtain the parameters of the timing trading rule:

[0086] min f CBuy (α Buy ,β Buy (31)

[0087] max f CSell (α Sell ,β Sell (32) The specific process of the Monte Carlo method is as follows: a large number of carbon price scenarios are randomly generated. Considering only carbon trading, simulations are performed daily for both buying and selling scenarios, starting from the randomly selected initial trading day and continuing until one carbon transaction is completed according to the corresponding timing trading rules. The average expected purchase price and expected selling price obtained from all scenarios are f respectively. CBuy (α Buy ,β Buy ) and f CSell (α Sell ,β Sell Differential evolution algorithm is a commonly used method well-known to those in the fields of operations research and optimization and the power system industry, and will not be described in detail here.

[0088] Before performing optimization, the carbon market one-way trading rules are directly added to the two-stage model, and the timing trading rules are added according to the following steps.

[0089]

[0090]

[0091] The final decision-making process for the power generator's daily electricity-carbon market throughout the year is as follows: Figure 3 As shown.

[0092] Step 4: Solve the optimization problem to obtain the decisions for the electricity market and carbon trading market, and conduct market transactions based on the decisions.

[0093] In some embodiments of the present invention, the effectiveness of the method is verified through simulations of numerous random scenarios. Regarding market prices, although the current proportion of free carbon allowances in China is relatively high and the carbon price is relatively low, in the long run, the proportion of free allowances will decrease while the carbon price will increase. Therefore, this embodiment sets up two environments: low carbon price and high carbon price. The low carbon price environment uses the 2022 national carbon market price data from the Shanghai Environment and Energy Exchange to estimate the OU process parameters, while the high carbon price model uses the 2022 EU ETS price data converted to RMB at a ratio of 1:7.39 to estimate the OU process parameters. In the example, the 2022 Guangdong spot market weighted average price data for power generation across the province is superimposed with a 10% normal distribution bias to generate the electricity price scenario and predict the electricity price. Missing values ​​in the above price data are filled with nearest-neighbor data, and the generated carbon price is controlled between 0 and the penalty price. Regarding cash flow, the maximum loan amount is 100 million yuan, the annual interest rate is 10%, the daily late payment fee rate is 5%, and deposit interest is ignored; daily expenses are generated using a normal distribution with a mean of 0 and a standard deviation of 10,000 yuan, and employee salaries of 5 million yuan are set to be paid on the 25th of each month; considering that the cash reserves of the generator may be insufficient at any time of the year, the cash reserves at the beginning of the year are set to 0 for the convenience of analysis.

[0094] To verify the effectiveness of the proposed method, this embodiment sets up the following four decision schemes and performs rolling optimization for 365 days a year under the condition of one-way carbon market trading.

[0095] Option 1: The method proposed in this invention. Considering cash flow constraints, a two-stage model is adopted for carbon allowance trading, which includes carbon market timing trading rules and carbon market one-way trading rules.

[0096] Option 2: No cash flow constraints. Based on Option 1, this allows for negative cash reserves, which is equivalent to not considering cash flow constraints.

[0097] Option 3: No cash flow constraints, no market timing rules. Negative cash reserves are allowed, and a two-stage stochastic decision-making method is used directly without adding market timing rules.

[0098] Option 4: No cash flow constraints, daily balance between emissions and quotas. This option allows for negative cash reserves and requires generators to accumulate net emissions equal to their daily quota holdings; this is a common practice in the literature.

[0099] There are also differences in the four schemes for processing the predicted carbon price for the aggregated month that is not extended to the day. After generating the carbon price curve with a specific day through the OU process, Schemes 1 and 2 use the timing trading rule to process the monthly carbon price, Scheme 3 uses the lowest monthly carbon price as its monthly carbon price, and Scheme 4 uses the average monthly carbon price as its monthly carbon price, so that the aggregated monthly carbon price is consistent with the decision-making strategy of each method.

[0100] This embodiment assumes that the carbon price curve cannot be accurately predicted; therefore, it uses the Ornstein-Uhlenbeck mean regression process (OU process) to establish a stochastic model for future carbon prices. The OU process is represented by a stochastic differential equation:

[0101]

[0102] Where t represents consecutive time intervals, B t For standard Brownian motion, Let be the carbon price at time t. Let η be the mean, η be the mean regression rate, and σ be the standard deviation. The carbon price is the observed value at time t0.

[0103] In both low and high carbon price environments, 100 scenarios were randomly generated for the entire year and then optimized daily for 365 days (the first scenario used actual data from 2022). During the daily optimization, 50 predicted scenarios were generated. The results of each scheme are shown in Tables 1 and 2. The estimated carbon cost is the average annual carbon purchase cost obtained on the first day of optimization.

[0104] Table 1. Mean results of 100 scenarios under low carbon price

[0105]

[0106] Table 2. Mean results of 100 scenarios under high carbon prices

[0107]

[0108] Comparing Tables 1 and 2, it can be seen that with the significant increase in carbon prices, the willingness of thermal power generators to generate electricity has been affected, resulting in a substantial decrease in their annual power generation, carbon emissions, and revenue. At the same time, the differences in effectiveness between different decision-making schemes have become more pronounced.

[0109] Comparing Schemes 1 and 2 reveals the impact of cash flow constraints proposed in this invention. Scheme 2, by not considering cash flow constraints, allows for unrestricted carbon purchases and power generation even with low cash reserves, resulting in higher returns, lower actual carbon costs, and greater total power generation and emissions compared to Scheme 1. However, Scheme 2 can lead to negative cash reserves, which is impractical in real-world scenarios.

[0110] A comparison of schemes 2, 3, and 4 reveals the advantages of the carbon trading method proposed in this invention. Scheme 2 achieves the highest returns and lowest carbon costs under the constraint of timing-based trading rules. Scheme 3 employs a two-stage model without timing-based trading rules, but the second stage always purchases carbon when the predicted carbon price is lowest, leading to an underestimation of carbon purchase costs. This results in Scheme 3 having higher actual carbon costs than Scheme 2, but higher total power generation and emissions, resulting in a loss of revenue. Scheme 4 requires that the daily purchased allowances equal the daily emissions, without choosing the timing of carbon purchases. Therefore, it has the highest average carbon purchase cost, the lowest total power generation and emissions, and lower returns. This reflects the necessity of considering annual carbon allowance settlement.

[0111] In some embodiments of the present invention, a power generator carbon market decision-making system considering cash flow is provided for implementing the methods in the foregoing embodiments, the system comprising the following modules:

[0112] The decision model building module is used to build a stochastic decision model for the electricity-carbon market with a full-year perspective and cash flow constraints.

[0113] A simplification module is used to reduce the number of decision periods by adopting mixed time step modeling and to simplify the annual view stochastic decision-making model of the electricity-carbon market into a two-stage stochastic optimization model.

[0114] The rule-building module is used to add carbon market one-way trading rules and carbon market timing trading rules to the annual view electricity-carbon market stochastic decision-making model. The carbon market one-way trading rule assumes that generators have a risk-averse tendency in the carbon market, do not speculate in the carbon trading market, and only conduct one-way trading based on future expectations. The carbon market timing trading rule is to buy and sell carbon allowances based on market carbon price, purchase price and selling price.

[0115] The solver module is used to perform the solution and obtain the decisions for the electricity market and carbon trading market.

[0116] The system in this embodiment corresponds to the method disclosed in the foregoing embodiments, so the description is relatively simple. For relevant details, please refer to the method section.

[0117] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A power generator carbon market decision-making method considering cash flow, characterized in that, Including the following steps: Step 1: Establish a stochastic decision-making model for the electricity-carbon market with a full-year perspective and cash flow constraints; specifically including: Without considering the settlement cycle, the daily optimization decision objective function for power generators is: In the formula, Represents the mathematical expectation. , d It is an index for each decision-making day within the year. This is the index for the current decision date. D For year-end index, For the deadline The gains already achieved by [date] , , , , They are respectively d Daily electricity market revenue, generation costs, carbon market carbon purchase expenditures, loan repayments, and daily operating expenses. This is the daily discount rate. Fines incurred for failing to pay emission quotas; yes d A vector of random variables for each day. for d Daily expenses that need to be paid; yes d Daily decision variable vector; After considering the settlement cycle, the daily optimization decision objective is set as maximizing the expected total assets of the distributor at the end of the year: In the formula, It's the cash I had at the end of the year. The electricity bill had not arrived by the end of the year. and It refers to outstanding loans at the end of the year and their penalty coefficients. and It refers to the remaining late payment amount and its penalty coefficient at the end of the year; Power generators earn profits by selling electricity in the electricity market. Generating electricity requires consuming fuel and must also comply with the operational constraints of the generating units. In the formula, For power generation d The electricity cost corresponding to daily power generation The duration is 1 hour. for d day h Electricity price at that time for d day h hour g The generating capacity of the unit For power generation d The cost of fuel consumed by daily power generation for g The cost coefficient of the unit and They are respectively g upper and lower power limits of the unit and They are respectively g Maximum uphill / downhill gradient per hour for the generator unit; Greenhouse gas emissions are generated during power generation. Under the carbon trading mechanism, power generators need to purchase allowances for their emissions and settle these allowances in full the following year; otherwise, they will face penalties. The expression is as follows: In the formula, For power generation d Net carbon emissions per day after deducting the free rate. and They are respectively g Emissions per unit of electricity generated by the unit and the amount of free allowance determined by the baseline method. for d The total amount of emission allowances held by Rifa Electric. for d The amount of carbon credits traded by Rifa Electric in the carbon market. From the beginning of the year to d The total net emissions generated by daily e-commerce. yes d Net emissions generated by daily e-commerce platforms. for d Fees incurred from daily trading quotas for d Daily carbon allowance price Fines incurred for failing to pay emission allowances The penalty for the unit's unpaid quota. This represents the total annual carbon emissions. Carbon allowances held by power generators at the end of the year; Establish a cash flow model for generators: In the formula, for d Daily cash reserves The daily interest rate for deposits. Is d The set of power generation dates for which daily electricity bills are payable. In order to be in d The set of power generation dates for which fuel costs are paid daily. In advance Application submitted on the day and d Loans that arrive within one day This is used to indicate the time required for loan application approval. for d Loans repaid daily for d Overdue payments due on the day for d Daily expenses for d Daily cumulative delayed payment amount This refers to the daily late payment fee rate. for d The total amount of outstanding loans on that day. The daily interest rate for the loan. This is the maximum loan amount for the company. for d The minimum repayment amount for a daily loan. This is the longest loan period in a single transaction. for d The minimum repayment amount for a daily loan. For the shortest period of a single loan, for The electricity cost corresponding to daily power generation for Fuel costs corresponding to daily power generation For the index of trading days to be considered, for Loans applied for on the same day for The corresponding loan application was approved d The number of days that have passed. for Loans applied for on the same day, for Loans applied for on the same day; Step 2: Use mixed time step modeling to reduce the number of decision periods and simplify the annual view of the stochastic decision-making model of the electricity-carbon market into a two-stage stochastic optimization model; Step 3: Add carbon market one-way trading rules and carbon market timing trading rules to the annual perspective electricity-carbon market stochastic decision-making model. The carbon market one-way trading rule assumes that generators have a risk-averse tendency in the carbon market, do not speculate in the carbon trading market, and only conduct one-way trading based on future expectations. The carbon market timing trading rule is to buy and sell carbon allowances based on market carbon price, purchase price and selling price. Step 4: Solve the problem to obtain the decisions for the electricity market and carbon trading market.

2. The power generation carbon market decision-making method considering cash flow as described in claim 1, characterized in that, In step 2, the two-stage stochastic optimization model only distinguishes between the present and future stages: in, , They represent from d 0 to D Decision variables and random variables for each day, f The objective function is denoted as .

3. The power generation carbon market decision-making method considering cash flow as described in claim 2, characterized in that, The two-stage stochastic optimization model was processed using a scenario-based approach, resulting in: in, N This represents the total number of scenes. n For scene indexing, For the scene n From d 0+1 days to D day Decision variables, For the scene n From d Day 0 to D Decision variables for the day For the scene n From d Day 0 to D The value of the random variable for the day.

4. The power generation carbon market decision-making method considering cash flow as described in claim 1, characterized in that, In the one-way carbon trading rules of the carbon market, only the shortage quotas are bought or the excess quotas are sold in the same scenario. There will be no buying first and then selling or selling first and then buying.

5. The power generation carbon market decision-making method considering cash flow as described in claim 1, characterized in that, The expression for the one-way carbon trading rules in the carbon market is: in, It is a 0-1 integer variable. 1 indicates the first n In one scenario, the carbon quota trading direction is buying; otherwise, it is selling. For the scene n Down d Daily carbon quota trading volume M Sell and M Buy These are constants, representing the maximum possible future quota selling volume and buying volume, respectively.

6. Power generator considering cash flow and carbon allowance settlement cycle as described in any one of claims 1-5 Market decision-making method, characterized by, The carbon market timing trading rules take the following form: (1) The market carbon price is less than or equal to the purchase price Carbon allowances can be purchased at any time; (2) The market carbon price is greater than or equal to the selling price. Carbon allowances can be sold at any time; (3) The market carbon price is at the purchase price and selling price Carbon quota trading will not be conducted during this period; (4) Carbon allowances can be freely bought and sold on the last trading day; Right now, d Daily carbon trading purchase price and selling price for In the formula, , , , These are parameters to be determined.

7. A power generator carbon market decision-making method considering cash flow as described in claim 6, characterized in that, A stochastic model of future carbon prices is established using the OU process, which is represented by a stochastic differential equation: In the formula, t Indicates consecutive moments, For standard Brownian motion, Let be the carbon price at time t. The mean, For the mean regression rate, Standard deviation, for t The observed carbon price at 0.

8. A power generator carbon market decision-making method considering cash flow as described in claim 6, characterized in that, Solve the following problem to obtain the parameters of the timing trading rule: In the formula, and These represent the expected purchase price and the expected sale price.

9. A power generator carbon market decision-making system that considers cash flow, characterized in that, The system for implementing the method of any one of claims 1-8 includes the following modules: The decision model building module is used to build a stochastic decision model for the electricity-carbon market with a full-year perspective and cash flow constraints. A simplification module is used to reduce the number of decision periods by adopting mixed time step modeling and to simplify the annual view stochastic decision-making model of the electricity-carbon market into a two-stage stochastic optimization model. The rule-building module is used to add carbon market one-way trading rules and carbon market timing trading rules to the annual view electricity-carbon market stochastic decision-making model. The carbon market one-way trading rule assumes that generators have a risk-averse tendency in the carbon market, do not speculate in the carbon trading market, and only conduct one-way trading based on future expectations. The carbon market timing trading rule is to buy and sell carbon allowances based on market carbon price, purchase price and selling price. The solver module is used to perform the solution and obtain the decisions for the electricity market and carbon trading market.

Images