Electricity-carbon-green evidence fused multi-market collaborative service balance decision-making method

Abstract

The invention relates to an electricity-carbon-green evidence fusion multi-market collaborative service equilibrium decision-making method. The method comprises the following steps: step 1, analyzing electricity-carbon-green evidence benefits under a multi-coupling mechanism; step 2, based on the electricity-carbon-green evidence benefit analysis result under the multi-coupling mechanism in the step 1, constructing a system operation optimization model under the multi-coupling mechanism; step 3, training and calibrating the system operation optimization model under the multi-coupling mechanism constructed in the step 2; 4, performing optimization verification on the system operation optimization model under the multi-coupling mechanism after training and calibration in the step 3; and step 5, using the system operation optimization model under the multi-coupling mechanism after optimization verification in the step 4 to output a daily optimal decision strategy to participate in market decision execution. According to the method, collaborative optimization of electric quantity distribution, carbon emission reduction accounting and green certificate configuration can be realized, and the green power consumption accuracy and the multi-market collaborative operation efficiency are improved.

Description

Technical Field

[0001] This invention belongs to the field of energy management technology and relates to a market entity service equilibrium decision-making method, especially a multi-market collaborative service equilibrium decision-making method integrating electricity, carbon and green certificates. Background Technology

[0002] With the gradual advancement of the global energy transition and the "dual carbon" goals, the coordinated development of the electricity market, carbon market, and green certificate market has become a key path to promote the development of sustainable energy systems. However, the current independent operation of these three markets has exposed a series of technical and practical problems, such as... Figure 1 The diagram illustrates the process of benefit analysis for multiple coupling mechanisms, including the electricity market, carbon market, and green certificate market. It details how benefit calculations for each market are performed through data collection and analysis, providing data support for subsequent optimization model construction. Specifically: Standardization issues in data collection: Currently, data from the electricity market, carbon market, and green certificate market are typically collected separately by different departments and platforms. Electricity market data is provided by electricity trading centers, usually at 15-minute intervals; carbon market data comes from carbon trading platforms, with daily or monthly cycles; and green certificate market data is aggregated based on green electricity generation. This inconsistency in data formats and time scales makes cross-market data fusion and analysis extremely difficult, especially when performing multi-market collaborative optimization, as it fails to guarantee data consistency and comparability. Therefore, existing methods face significant technical challenges in integrating multi-market data, often leading to substantial biases in decision-making outcomes.

[0003] Limitations of model building: Traditional optimization models often only consider the economics or emission reduction costs of a single market. For example, some models focus solely on the supply and demand balance of the electricity market, neglecting the impact of the carbon and green certificate markets on the system. This model structure not only fails to effectively reflect the interactions between different markets but also fails to optimize the overall efficiency of resource allocation. In such cases, insufficient coordination among the markets leads to the inability to achieve optimal system operation. Existing single-layer models cannot effectively capture the dynamic interactions between the electricity, carbon, and green certificate markets, resulting in low efficiency in resource allocation and utilization.

[0004] Neglecting the load characteristics of industrial users and policy constraints: Currently, many existing methods fail to differentiate between different industrial users (such as energy-intensive industries and export-oriented manufacturing) when dealing with their load characteristics. These industrial users exhibit significantly different load characteristics; for example, energy-intensive industries typically have continuous production loads, while export-oriented manufacturing usually exhibits cyclical loads. Existing models fail to accurately account for these differences, leading to significant discrepancies between optimization results and actual demand. Furthermore, existing methods fail to effectively translate policy constraints (such as Renewable Energy Consumption Scheme (RPS) and carbon quota allocation) into mathematical constraints, potentially resulting in optimization results that violate relevant policy requirements and lack compliance.

[0005] In summary, existing technologies suffer from drawbacks such as inconsistent data, significant model limitations, and insufficient consideration of load characteristics and policy constraints. There is an urgent need for a multi-market collaborative service equilibrium decision-making method that integrates electricity, carbon, and green certificates to achieve coordinated optimization of the electricity market, carbon market, and green certificate market, thereby realizing effective collaboration among multiple markets and optimal allocation of resources.

[0006] A search revealed no prior art patents that are identical or similar to this invention. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention proposes a multi-market collaborative service equilibrium decision-making method that integrates electricity, carbon, and green certificates. This method can achieve optimal decomposition of grid-connected power in an aggregated system of industrial users and electricity sales companies under multiple coupling mechanisms by using the revenue and transaction prices of the electricity market, carbon market, and green certificate market. It can accurately match the green electricity consumption responsibilities and carbon emission reduction needs of market participants, thereby improving the accuracy of multi-market collaborative operation and green electricity consumption.

[0008] The above-mentioned objective of this invention is achieved through the following technical solution: A multi-market collaborative service equilibrium decision-making method integrating electricity, carbon, and green certificates includes the following steps: Step 1: Analyze the benefits of electricity-carbon-green certificates under multiple coupling mechanisms; Step 2: Based on the results of the electricity-carbon-green certificate benefit analysis under the multi-coupling mechanism in Step 1, construct a system operation optimization model under the multi-coupling mechanism; Step 3: Train and calibrate the system operation optimization model under the multi-coupling mechanism constructed in Step 2; Step 4: Optimize and verify the system operation optimization model under the multi-coupling mechanism after training and calibration in Step 3; Step 5: Using the optimized and verified multi-coupling mechanism system operation optimization model from Step 4, output the daily optimal decision-making strategy and participate in market decision execution.

[0009] Furthermore, the specific steps of step 1 include: 1.1 Profit Analysis of Participating in the Electricity Market under Multiple Coupling Mechanisms: The formula for calculating electricity market revenue is as follows: in, This represents the output of the m-th photovoltaic unit at time t. , These represent the charging and discharging power of the nth energy storage power station at time t; The load demand produced by industrial users at time t; The price is the spot market price; M is the number of photovoltaic units; L is the number of wind turbines; N is the number of energy storage units. Relc represents the carbon balance value of the nth wind turbine output at time t.

[0010] 1.2 Profit Analysis of Participating in the Carbon Market under Multiple Coupling Mechanisms: Carbon market revenue calculation formula: in, This refers to the total amount of carbon dioxide emissions offset by new energy power generation; , These refer to the on-grid electricity generated by photovoltaic power generation and wettable generator sets for the approved CCER (China Certified Electricity Recycling) units; Coal consumption rate of coal-fired power units; The standard coal carbon emission factor; The carbon dioxide emission factor per unit carbon combustion is 3.67.

[0011] 1.3 Profit Analysis of Participating in the Green Certificate Market under Multiple Coupling Mechanisms: Green certificate market return calculation formula: in, It is revenue from the green certificate market; It is the total feed-in power for which photovoltaic green certificates have been approved; It is the price of a green certificate for photovoltaic power generation; It is the total feed-in power of the approved wind power green certificate; It is the total feed-in power of the approved wind power green certificate; The revenue of the system under the multi-coupling mechanism of participating in the carbon market is: in, It is income from participating in the carbon market. This is the trading price of a CCER certificate.

[0012] The total benefit for industrial users and electricity sales companies participating in the "electricity-carbon-green certificates" multi-market is the sum of the benefits from all three markets, namely: Furthermore, the specific steps of step 2 include: 2.1 Constructing the upper-level model: The upper-level model is based on the principles of "self-use priority" and "surplus power grid connection" to optimize the charging and discharging status of ES stations within the system. By optimizing the power declaration for the system to participate in the electricity market, it improves the matching degree between grid-connected power and industrial load.

[0013] The upper-level model needs to meet constraints such as unit output, energy storage output, energy storage status, and energy storage capacity.

[0014] The output constraints of new energy units are as follows: Load coverage constraints: For energy storage charging and discharging power constraints, the energy storage charging and discharging power must be less than the maximum charging and discharging power: in, , It is the minimum charging and discharging power of the nth energy storage unit; , It is the maximum charge and discharge power of the nth energy storage unit; , These are the discharge and charge state variables of the nth ES unit, respectively.

[0015] Due to limitations in energy storage operation, charging and discharging cannot occur simultaneously. For energy storage capacity constraints, the energy storage capacity shall not exceed the upper or lower limit of the rated capacity; the capacity at the next moment shall be determined by the remaining capacity at the previous moment and the charging and discharging behavior at that moment.

[0016] To ensure the parallelism of scheduling, the capacity of the initial state and the final state within a cycle are equal; in, It is the capacity of the nth energy storage unit at time t; , These refer to the charging and discharging efficiencies of energy storage, respectively. , These represent the energy storage capacity at the initial and final moments, respectively.

[0017] 2.2 Constructing the lower-level model: The objective function is to maximize the sum of the returns from the carbon market and the green certificate market. Energy storage power stations experience energy losses during charging and discharging. GC and CCER verifications are based on online electricity consumption; therefore, the verified electricity consumption is the actual power generation minus the energy losses. The ratio of photovoltaic (PV) to wind power in energy storage, charging, and discharging is determined based on the prevailing ratio of PV to wind turbine power generation.

[0018] in, It represents the percentage of photovoltaic unit output within time t.

[0019] Furthermore, the specific steps of step 3 include: 3.1 Data Collection and Standardization: Collect load data, electricity market transaction data, carbon market data, and green certificate market data from industrial users over the past three years; divide the training set and test set according to different time scales and states, and standardize all input features; 3.2 Training of the two-layer optimization model: (1) First, based on the collected data and policy requirements, set the initial values ​​of the upper-level decision variables; (2) Take the initial decision of the upper layer as the hard constraint of the lower layer model, and solve the local optimal solution of the lower layer, that is, the optimal operation strategy of the "follower" under the rules given by the "leader". (3) Substitute the solution results of the lower layer into the upper layer model, modify the decision variables of the upper layer, and realize the adaptive adjustment of the "leader" to the response of the "follower": repeat steps (2)-(3) until the following convergence conditions are met, and stop the iteration.

[0020] Furthermore, the specific method for step 4 is as follows: After training converges, the effectiveness of the model needs to be ensured through multi-dimensional validation, and key parameters need to be calibrated. Feasibility verification: Check whether the training results satisfy all constraints. If there are constraints that are violated, backtrack and adjust the initial values ​​of the upper-level decision variables or the constraint weights. Optimality verification: Compare the training results with different initial values ​​to confirm that the final converged upper-level objective function value is globally optimal, thus avoiding optimization traps caused by initial value bias; Robustness verification: Perturb the input data, retrain the model, and observe the change in the value of the upper objective function. Otherwise, the elasticity coefficient of the model constraint needs to be adjusted.

[0021] Furthermore, step 5, where the model runs and outputs the daily optimal decision-making strategy, includes: Electricity allocation between medium- and long-term contracts and the spot market in the electricity market; Recommendations on CCER development and quota trading in the carbon market; Green certificate purchase or sale strategies in the green certificate market to meet renewable energy consumption responsibility (RPS) requirements.

[0022] The advantages and beneficial effects of this invention are as follows: 1. This invention proposes a multi-market collaborative service equilibrium decision-making method integrating electricity, carbon, and green certificates. For the first time, it incorporates electricity price signals from the electricity market, carbon quota prices / CCERs from the carbon market, and environmental premiums from the green certificate market into a unified optimization framework, achieving dynamic interaction among the three through a two-layer model. By utilizing the revenue and transaction prices from the electricity, carbon, and green certificate markets, a two-layer optimization model adapted to the needs of multiple industries is constructed, resolving issues such as misalignment of stakeholders, data disconnect, and demand bias in existing models. This achieves optimized allocation of grid-connected electricity and precise matching of green electricity consumption under multiple coupling mechanisms.

[0023] 2. This invention provides a multi-market collaborative service equilibrium decision-making method integrating electricity, carbon, and green certificates. It abandons the traditional model's focus on "rural users" and instead focuses on industrial users (high-energy-consuming industries, export-oriented manufacturing, and high-precision manufacturing) and electricity sales companies, aligning with the reality in the technical report that "market players are mainly large enterprises," thus solving the problem of scenario misalignment. Furthermore, it is the first to incorporate the "medium-to-long-term + spot" electricity market, the "quota + CCER" carbon market, and the "RPS compliance + trading" green certificate market into a unified framework. Through a two-layer model, it achieves closed-loop optimization of "electricity allocation → carbon-green certificate response → revenue feedback," avoiding the limitations of single-market optimization. In addition, it closely integrates with national policy requirements, transforming policy indicators into model constraints to ensure that the optimization results not only comply with policies but also improve the operational performance of the electricity-carbon-green certificate collaborative configuration and the accuracy of green electricity consumption, reducing the uncertainty of compliance costs.

[0024] 3. The multi-market collaborative service equilibrium decision-making method integrating electricity, carbon, and green certificates proposed in this invention aims to solve the problems exposed in the background technology. Through the following key steps, this invention effectively overcomes the defects in the prior art and realizes the optimization of the system operation state and the coordination between green electricity, carbon emission reduction, and green certificate compliance under multi-market collaboration. From the initial stage of electricity, carbon, and green certificate market benefit analysis under multiple coupling mechanisms, to the construction, training, and verification of the system optimization model under multiple coupling mechanisms, the optimal decision of the system is finally achieved.

[0025] (1) Analysis of the benefits of electricity-carbon-green certificates under multiple coupling mechanisms: Regarding data acquisition and standardization, this invention integrates data from the electricity, carbon, and green certificate markets through unified data standardization processing, ensuring that all data are on the same time scale (e.g., 48 hours). This solves the problems of inconsistent data formats and mismatched time scales in traditional methods. Through this data standardization process, this invention can provide consistent and reliable input data for subsequent model building, making collaborative optimization across multiple markets possible.

[0026] (2) Construction of the system operation optimization model: This invention constructs a two-layer optimization model to achieve closed-loop optimization of "electricity allocation → carbon-green certificate response → feedback adjustment". The upper-layer model mainly focuses on the electricity declaration for system participation in the electricity market, optimizing the declaration through the principles of "self-consumption priority" and "surplus electricity grid connection" to improve the coordination between industrial user-side energy consumption and grid-connected electricity. The lower-layer model optimizes the electricity allocation for system participation in the carbon market and green certificate market to improve the matching degree between carbon emission reductions, green certificate holdings, and policy quotas. Compared with traditional single-market optimization methods, this invention can achieve coordinated optimization between multiple markets, overcoming the inefficiency problems caused by independent optimization in each market in existing technologies, and improving the system's operating efficiency and green electricity consumption capacity in multi-market collaborative participation scenarios.

[0027] (3) Model training and calibration: By training and calibrating models based on historical data, this invention not only meets the policy requirements of the electricity market, carbon market, and green certificate market, but also improves the accuracy of the model's coordinated response to electricity, carbon, and green certificates in different regions and user types. During training, all input data undergoes standardization to ensure model consistency and generalization ability across different regional dimensions. Input data and decision variables for each market are appropriately adjusted within the model to ensure that the optimization results not only meet policy compliance requirements but also improve the accuracy of electricity allocation, carbon emission reduction accounting, and green certificate configuration.

[0028] (4) Optimize model validation: This invention employs multi-dimensional verification methods, including feasibility verification, optimality verification, and robustness verification, to ensure the stability and reliability of the optimization model in practical applications. Feasibility verification checks whether the model satisfies all constraints, ensuring the optimization results are legal and compliant. Optimality verification compares training results with different initial values ​​to ensure the model eventually converges to the global optimum. Robustness verification tests the model's stability when facing disturbance inputs, assessing the model's sensitivity to external disturbances by observing changes in the upper-level objective function (such as comprehensive indicators like grid-connected power allocation deviation and green energy consumption deviation). Through this series of verifications, this invention effectively improves the stability and applicability of the decision-making strategy, ensuring the model's efficiency and reliability in practical operation.

[0029] (5) Model application and market decision execution: The core value of this invention lies in transforming the optimized model into actual market operation and dispatch instructions. By deploying the model to actual business systems such as power trading, carbon quota management, and green certificate issuance, and constructing a decision support system (DSS), this invention can automatically generate executable power allocation, trading, and dispatch strategies based on real-time market data (such as renewable energy output forecasts, spot electricity prices, carbon prices, and green certificate prices). The system can automatically or assisted users in executing transactions on the corresponding market platform and respond to dispatch based on the real-time status of the power grid, achieving precise control over energy storage charging and discharging, renewable energy output, etc. Simultaneously, the system provides operation monitoring and dynamic adjustment functions to ensure the adaptability of the strategy under market fluctuations and automatically generates compliance reports to support policy audits. Ultimately, this invention achieves a complete closed loop from "model optimization" to "market operation," improving the decision-making efficiency, green electricity consumption capacity, and compliance assurance level of market participants in a multi-market collaborative environment.

[0030] (6) Implementation of the optimal decision of the system: This invention can precisely match the needs of industrial users and electricity sales companies in the integrated trading of electricity, carbon, and green certificates, ensuring coordination between green electricity consumption responsibilities, carbon emission reduction requirements, and the collaborative operation of multiple markets. Ultimately, through multi-market collaborative optimization, this invention achieves optimization of system operation and resource allocation, and provides industrial users and electricity sales companies with decision-making strategies that can be directly implemented in engineering practice, overcoming the problem that existing methods in the background art cannot take into account multiple markets and different user types. Attached Figure Description

[0031] Figure 1 This is a benefit analysis diagram under the multiple coupling mechanisms of the present invention; Figure 2 This is a flowchart illustrating the model construction process of the present invention. Figure 3 The flowchart of the multi-coupling mechanism of the multi-market collaborative service equilibrium decision-making method integrating electricity, carbon and green certificates of this invention is shown. Detailed Implementation

[0032] The embodiments of the present invention will be further described in detail below with reference to the accompanying drawings: A multi-market collaborative service equilibrium decision-making method integrating electricity, carbon, and green certificates, such as... Figures 1 to 3 As shown, the package includes the following steps: Step 1: Analyze the benefits of electricity-carbon-green certificates under multiple coupling mechanisms; The specific steps of step 1 include: 1.1 Profit Analysis of Participating in the Electricity Market under Multiple Coupling Mechanisms: We collect and analyze renewable energy forecasts and industrial user load data (continuous production load of high-energy-consuming industries, periodic load of export-oriented manufacturing, etc.) on a 48-hour timescale, adapting them to industrial users' production plans and the electricity market under different mechanisms.

[0033] The electricity market comprises medium- and long-term contracts and spot market transactions: industrial users and electricity retailers purchase green electricity based on medium- and long-term contracts, while simultaneously balancing load deviations through the spot market. In mature spot markets, electricity prices fully reflect real-time supply and demand, and system operating objectives are highly consistent with the electricity market's balancing needs.

[0034] The formula for calculating electricity market revenue is as follows: in, This represents the output of the m-th photovoltaic unit at time t. , These represent the charging and discharging power of the nth energy storage power station at time t; The load demand produced by industrial users at time t; The price is the spot market price; M is the number of photovoltaic units; L is the number of wind turbines; N is the number of energy storage units. It is the output of the nth wind turbine at time t. Carbon payoff balance.

[0035] 1.2 Profit Analysis of Participating in the Carbon Market under Multiple Coupling Mechanisms: Industrial users can reduce carbon allowance consumption through green electricity consumption or supplement carbon allowance gaps by developing CCERs, thereby realizing carbon market benefits. The green attributes of wind and solar power units can benefit from China Certified Emission Reductions (CCERs) in the carbon market. One CCER is equivalent to one carbon allowance. Aggregators select to allocate electricity based on trading prices in the green certificate market and the carbon market, and approve GCs and voluntary carbon emission reduction allowances. Electricity with approved GCs cannot be further approved for carbon emission reductions.

[0036] The effective online electricity consumption of new energy sources is equivalent to offsetting thermal power generation. Based on thermal power output, standard coal consumption, and carbon emission factor of standard coal combustion, the reduction in carbon dioxide emissions from new energy grid connection is calculated.

[0037] Carbon market revenue calculation formula: in, This refers to the total amount of carbon dioxide emissions offset by new energy power generation; , These refer to the on-grid electricity generated by photovoltaic power generation and wettable generator sets for the approved CCER (China Certified Electricity Recycling) units; Coal consumption rate of coal-fired power units; The standard coal carbon emission factor; The carbon dioxide emission factor per unit carbon combustion is 3.67.

[0038] 1.3 Profit Analysis of Participating in the Green Certificate Market under Multiple Coupling Mechanisms: Green certificates are core credentials for industrial users to fulfill their Renewable Energy Consumption Responsibility (RPS) and prove their green electricity consumption. Industrial users are required to hold sufficient green certificates according to policy requirements, and can supplement gaps or sell surpluses through green certificate trading. Wind power and solar power are clean energy sources with green and environmentally friendly attributes. GCs are issued based on online electricity consumption. Electricity generated using GCs is traded in the GC trading market, generating economic benefits and receiving green electricity subsidies through GC sales. Actual online electricity consumption under different mechanisms is collected from smart meters, and GCs are issued and traded based on this online consumption.

[0039] Green certificate market return calculation formula: in, It is revenue from the green certificate market; It is the total feed-in power for which photovoltaic green certificates have been approved; It is the price of a green certificate for photovoltaic power generation; It is the total feed-in power of the approved wind power green certificate; It is the total feed-in power of the approved wind power green certificate; The revenue of the system under the multi-coupling mechanism of participating in the carbon market is: in, It is income from participating in the carbon market. This is the trading price of a CCER certificate.

[0040] The total benefit for industrial users and electricity sales companies participating in the "electricity-carbon-green certificates" multi-market is the sum of the benefits from all three markets, namely: Step 2: Construction of system operation optimization model under multiple coupling mechanisms; With the goal of achieving comprehensive benefits across multiple markets, a two-layer model is used to optimize resource allocation in the electricity market, carbon market, and green certificate market, clarifying the dynamic coupling relationships between these markets. Under the premise of ensuring load supply, equipment operating boundaries, and policy constraints, the system achieves coordinated optimization among grid-connected electricity, carbon emission reductions, and green certificate allocation, thereby improving the accuracy of green electricity consumption and system operating efficiency.

[0041] The specific steps of step 2 include: With the goal of improving the accuracy of electricity allocation and green energy consumption, a two-layer model is used to optimize the system's resource allocation in the electricity market, carbon market, and green certificate market, clarifying the dynamic coupling relationship between the various markets: 2.1 Upper-level model: The upper-level model optimizes the electricity declaration process for the system's participation in the electricity market to improve the coordinated allocation of energy consumption between industrial users and grid-connected electricity. The objective function is: The upper-level model needs to meet constraints such as unit output, energy storage output, energy storage status, and energy storage capacity.

[0042] The output constraints of new energy units are as follows: Load coverage constraints: For energy storage charging and discharging power constraints, the energy storage charging and discharging power must be less than the maximum charging and discharging power: in, , It is the minimum charging and discharging power of the nth energy storage unit; , It is the maximum charge and discharge power of the nth energy storage unit; , These are the discharge and charge state variables of the nth ES unit, respectively.

[0043] Due to limitations in energy storage operation, charging and discharging cannot occur simultaneously. For energy storage capacity constraints, the energy storage capacity shall not exceed the upper or lower limit of the rated capacity; the capacity at the next moment shall be determined by the remaining capacity at the previous moment and the charging and discharging behavior at that moment.

[0044] To ensure the parallelism of scheduling, the capacity of the initial state and the final state within a cycle are equal; in, It is the capacity of the nth energy storage unit at time t; , These refer to the charging and discharging efficiencies of energy storage, respectively. , These represent the energy storage capacity at the initial and final moments, respectively.

[0045] 2.2 Lower-level model: The lower-level model, building upon the electricity allocation results from the upper level, aims to improve the alignment between carbon emission reductions and green certificate holdings relative to policy quotas. The objective function is: Energy storage power stations experience energy losses during charging and discharging. GC and CCER verifications are based on online electricity consumption; therefore, the verified electricity consumption is the actual power generation minus the energy losses. The ratio of photovoltaic (PV) to wind power in energy storage, charging, and discharging is determined based on the prevailing ratio of PV to wind turbine power generation.

[0046] in, It represents the percentage of photovoltaic unit output within time t.

[0047] Step 3. Train and calibrate the model from Step 2; The specific steps of step 3 include: 3.1 Data Collection and Standardization: Load data from industrial users (categorized by industry: data from steel companies in high-energy-consuming industries, data from export-oriented manufacturing electronics companies, etc.), electricity market transaction data (medium- and long-term green electricity contract prices, spot prices, etc.), carbon market data (quota prices, CCER trading volume, etc.), and green certificate market data (issuance volume, trading prices, etc.) were collected over the past three years. Training and testing sets were divided according to different time scales and states. All input features were standardized to ensure the model's generalization ability and consistency across regional dimensions.

[0048] 3.2 Training of the two-layer optimization model: First, based on the collected data and policy requirements, initial values ​​for the upper-level decision-making variables are set; Secondly, the initial decision of the upper layer is used as a hard constraint of the lower layer model to solve the local optimal solution of the lower layer, that is, the optimal operation strategy of the "follower" under the rules given by the "leader". Finally, substitute the results from the lower layer into the upper layer model to correct the decision variables of the upper layer and achieve adaptive adjustment of the "leader" to the "follower" response: repeat steps 2-3 until the following convergence condition is met, and stop iterating (to avoid infinite loops or overtraining).

[0049] Step 4: Validate the optimized model from Step 3; The specific method for step 4 is as follows: After training converges, the effectiveness of the model needs to be ensured through multi-dimensional validation, and key parameters need to be calibrated. Feasibility verification: Check whether the training results satisfy all constraints. If there are constraints that are violated, backtrack and adjust the initial values ​​of the upper-level decision variables or the constraint weights. Optimality verification: Compare the training results with different initial values ​​to confirm that the final converged upper-level objective function value is globally optimal, thus avoiding optimization traps caused by initial value bias; Robustness verification: Apply random perturbations of ±5% to ±20% to the input data (including new energy output forecast series, electricity spot price series, industrial load curve, carbon price, green certificate price, etc.), retrain the model, and observe the changes in comprehensive indicators such as grid-connected power allocation deviation and green electricity consumption deviation in the upper objective function; when the change exceeds the preset threshold, adjust the elasticity coefficient or penalty weight of the model constraints to improve the model's adaptability to input perturbations.

[0050] Step 5: Implement market decisions using the model from Step 4; After the model is trained and validated, it will be deployed to the actual market environment to realize the coordinated operation and scheduling decision-making of multiple markets including "electricity-carbon-green certificates".

[0051] The specific method for step 5 is as follows: The trained two-layer optimization model is embedded into the power trading platform, carbon quota management system, and green certificate issuance and trading system to achieve real-time data interaction between the model and the market data platform. Develop a decision support system (DSS) to provide industrial users and electricity sales companies with a visual operating interface that supports real-time input of parameters such as load plans, market quotations, and policy constraints.

[0052] The system automatically acquires real-time data daily, including forecasts of renewable energy output for the next 48 hours, electricity spot market prices, carbon quota and CCER prices, green certificate trading prices, and user load plans.

[0053] Based on this, step 5, where the model runs and outputs the daily optimal decision-making strategy, includes: Electricity allocation between medium- and long-term contracts and the spot market in the electricity market; Recommendations on CCER development and quota trading in the carbon market; Green certificate purchase or sale strategies in the green certificate market to meet renewable energy consumption responsibility (RPS) requirements.

[0054] Based on the strategies generated by the model, the system automatically or semi-automatically executes transactions on the corresponding market platforms through interfaces. This includes submitting electricity declarations on the power trading platform, trading quotas or CCERs on the carbon trading platform, and completing green certificate subscriptions or resales on the green certificate platform. Simultaneously, the dispatch system executes real-time dispatch instructions based on the energy storage charging and discharging plans and renewable energy output curves output by the model, prioritizing "maximizing self-consumption and feeding surplus electricity into the grid."

[0055] Establish an operational monitoring dashboard to display key performance indicators such as system revenue, carbon emission reductions, and green certificate holdings in various markets in real time. If the market experiences significant fluctuations (such as sudden changes in electricity prices or temporary policy adjustments), the system will activate a re-optimization mechanism, rerunning the model based on the latest data and dynamically adjusting subsequent trading and scheduling strategies. The model output can automatically generate documents such as green electricity consumption certificates, carbon emission reduction accounting reports, and green certificate compliance lists, providing data support for enterprises to cope with government audits and carbon verifications, and providing a basis for enterprise operational scheduling optimization and compliance management.

[0056] The model outputs can automatically generate documents such as green energy consumption certificates, carbon emission reduction accounting reports, and green certificate compliance lists, providing enterprises with complete and reliable data support for responding to government audits and carbon verifications. The system regularly generates collaborative decision-making reports on "electricity-carbon-green certificates," providing a basis for enterprises' strategic decision-making and compliance management.

[0057] The working principle of this invention is: This invention proposes a multi-market collaborative service equilibrium decision-making method integrating electricity, carbon, and green certificates, comprising the following steps: ① Benefit analysis of electricity-carbon-green certificates under multiple coupling mechanisms, including: benefit analysis of participating in the electricity market under multiple coupling mechanisms: collect and analyze new energy forecast output data and industrial user load data (continuous production load of high energy-consuming industries, periodic load of export-oriented manufacturing, etc.) on a 48-hour time scale under different mechanisms to adapt to industrial user production plans and the electricity market.

[0058] Benefit analysis of participating in the carbon market under multiple coupling mechanisms: Data on carbon emission rights trading volume and carbon market prices are collected from the national carbon trading platform, as well as data on green electricity consumption and renewable energy output, and carbon emission factors are collected from the power trading center and the Ministry of Ecology and Environment. This is an important foundation for calculating emission reductions by replacing thermal power with green electricity.

[0059] Profit analysis of participating in the green certificate market under a multi-coupling mechanism: core data on green certificate issuance and trading are obtained from the national green certificate issuance and trading system, including the issuance volume of green certificates and the corresponding green electricity, green certificate trading price and trading volume. This allows for accurate calculation of the total amount of green certificates held and quantification of the direct benefits of green certificate trading.

[0060] ② Based on the analysis results, a system operation optimization model under multiple coupling mechanisms was constructed. To demonstrate the coupling relationship between the cluster system and various markets, a two-layer model was built: Upper-level model: Based on the principles of "maximizing self-consumption" and "grid connection of surplus power", the upper-level model optimizes the charging and discharging status of ES stations within the system and maximizes profits by optimizing the power declaration of the system to participate in the power market.

[0061] Lower-level model: The lower-level model optimizes the benefits of system clusters participating in the carbon market and green certificate market, and achieves the optimal allocation of electricity connected to the green certificate market and carbon market.

[0062] ③ Train the model based on the constructed model. Data collection and standardization: Collect load data from industrial users over the past 3 years (categorized by industry: data from steel companies in high-energy-consuming industries, data from export-oriented manufacturing electronics companies, etc.), electricity market transaction data (medium- and long-term green electricity contract prices, spot prices, etc.), carbon market data (quota prices, CCER trading volume, etc.), and green certificate market data (issuance volume, trading prices, etc.). Divide the training and test sets according to different time scales and states, and standardize all input features to ensure the model's generalization ability and consistency across regional dimensions.

[0063] Two-layer optimization model training: First, based on the collected data and policy requirements, set the initial values ​​of the upper-layer decision variables; second, use the upper-layer initial decisions as hard constraints for the lower-layer model, and solve for the local optimum of the lower layer, that is, the optimal operating strategy of the "followers" under the rules given by the "leader"; finally, substitute the solution of the lower layer into the upper-layer model, correct the upper-layer decision variables, and realize the adaptive adjustment of the "leader" to the response of the "followers": repeat steps ②-③ until the following convergence condition is met, and stop iterating (to avoid infinite loops or overtraining).

[0064] ④ Optimize model validation. After training convergence, the model's effectiveness needs to be ensured through multi-dimensional validation, and key parameters need to be calibrated: Feasibility verification: Check whether the training results satisfy all constraints. If there is a constraint violation, backtrack and adjust the initial values ​​of the upper-level decision variables or the constraint weights.

[0065] Optimality verification: By comparing the training results with different initial values, the final converged upper-level objective function value is confirmed to be globally optimal, thus avoiding optimization traps caused by initial value bias.

[0066] Robustness verification: Perturb the input data, retrain the model, and observe the change in the value of the upper objective function. Otherwise, the elasticity coefficient of the model constraint needs to be adjusted.

[0067] To further illustrate the feasibility of the decision-making method of this invention, a simple example is provided below. This example uses simplified parameters and is only used to demonstrate the modeling and solution process of the decision-making method of this invention. Those skilled in the art can scale up and generalize it based on actual engineering data.

[0068] Example: An industrial park participates in the coordinated operation of the electricity market, carbon market, and green certificate market within a single day (48 time slots, each lasting one hour). The system includes: One photovoltaic unit with a rated power of 0.8MW; One wind turbine unit with a rated power of 0.6MW; One energy storage device with a capacity of 2MWh and a maximum charging and discharging power of 1MW; The combined load of industrial users is 0.8 to 1.2 MW.

[0069] 1. Input data (example) Photovoltaic power forecast: The power output shows a single peak distribution during the daytime period from 10:00 to 16:00, with a maximum value of 0.6MW; Forecasted wind power output: fluctuating between 0.2 and 0.5 MW throughout the day; Industrial load: fluctuating between 0.8 and 1.2 MW; Electricity spot price Price: 0.20–0.45 yuan / kWh; CCER price 65 yuan / ton CO2; Green certificate prices: 180 yuan / certificate for photovoltaic power and 130 yuan / certificate for wind power. Energy storage efficiency: charging efficiency Discharge efficiency ; The baseline emission parameters for thermal power plants are: coal consumption rate φ, standard coal carbon emission coefficient σ, and carbon dioxide emission coefficient per unit of carbon combustion μ = 3.67, with values ​​taken according to current standards.

[0070] The above data was standardized and unified to a time scale of 48 time periods, which were then used as inputs to the models in steps 1 and 2.

[0071] 2. Solution results of the upper-level model (example) After executing the upper-level model of the decision-making method of this invention, the following result fragment is obtained within a typical day: At 10:00, the total output of photovoltaic and wind power is 1.0MW, and the industrial load is 0.8MW. The optimal strategy given by the model is: to charge 0.1MW of energy storage and send 0.1MW of grid-connected power to the grid. At 15:00, the total output of photovoltaic and wind power is 0.7MW, and the industrial load is 1.1MW. The optimal strategy given by the model is: energy storage discharges 0.3MW, and there is still 0.1MW of electricity to be purchased from the grid.

[0072] By solving 48 time periods, the upper-level model generated a complete power reporting and energy storage charging and discharging plan.

[0073] 3. Lower-level model solution and coordinated allocation of electricity, carbon, and green certificates Based on the grid-connected electricity and available renewable energy for certification given by the upper layer, the lower-layer model allocates the electricity available for CCER certification and green certificate issuance: Statistics show that the effective on-grid electricity from new energy sources throughout the day was 620 kWh; Of these, the effective power generation from photovoltaic power is 400 kWh, and the effective power generation from wind power is 220 kWh.

[0074] According to the carbon emission reduction calculation formula, we can obtain: After substituting typical parameters, the carbon emission reduction is approximately 2.034 tons of CO2.

[0075] The number of green certificates issued for photovoltaic power is approximately 0.4 (based on 1 certificate per 1000 kWh), while the number of green certificates issued for wind power is approximately 0.22.

[0076] The lower-level model, under the conditions of satisfying RPS constraints, carbon quota constraints, and green certificate market rules, provides: Prioritize the use of some renewable energy electricity for CCER certification to meet carbon emission reduction constraints; The remaining capacity is allocated for green certificate issuance to meet the green electricity consumption responsibilities of industrial users and their subsequent green certificate trading needs.

[0077] 4. Results Analysis By comparing the baseline strategy of "not adopting the decision-making method of this invention, but only allocating electricity based on experience," we can observe that: Under the same renewable energy output and load conditions, the grid-connected power curve obtained by the decision-making method of this invention is smoother during peak and valley periods, which is beneficial to improving the stability of system operation. The significant reduction in the discrepancy between carbon emission reductions, green certificate holdings, and policy quotas indicates that the decision-making method of this invention can more accurately match green electricity consumption responsibilities with carbon emission reduction requirements. After applying a ±10% perturbation to the prediction curve and price series and resolving the problem, the change in the objective function remained within the preset threshold, indicating that the model has good robustness.

[0078] This embodiment demonstrates that the decision-making method of the present invention can be directly executed in engineering scenarios, enabling optimized power allocation and precise control of green power consumption under the synergy of multiple markets including electricity, carbon, and green certificates.

[0079] It should be emphasized that the embodiments described in this invention are illustrative rather than limiting. Therefore, this invention includes, but is not limited to, the embodiments described in the specific implementation. Any other implementations derived by those skilled in the art based on the technical solutions of this invention are also within the scope of protection of this invention.

Claims

1. A multi-market collaborative service equilibrium decision-making method integrating electricity, carbon, and green certificates, characterized in that: The package comprises the following steps: Step 1, analyze the electricity-carbon-green certificate benefits under the multi-coupling mechanism; Step 2, based on the results of the electricity-carbon-green certificate benefit analysis under the multi-coupling mechanism in step 1, build a system operation optimization model under the multi-coupling mechanism; Step 3, train and calibrate the system operation optimization model under the multi-coupling mechanism built in step 2; Step 4, optimize and verify the system operation optimization model under the multi-coupling mechanism trained and calibrated in step 3; Step 5, use the system operation optimization model under the multi-coupling mechanism optimized and verified in step 4 to output daily optimal decision strategies to participate in market decision execution.

2. The method of claim 1, wherein the method is characterized by: The specific steps of step 1 include: 1.1 Analysis of the benefits of participating in the electricity market under the multi-coupling mechanism: The electricity market revenue calculation formula is as follows: ; wherein, is the mth photovoltaic unit output at time t, , is the charge-discharge power of the nth energy storage unit at time t; is the load demand produced by the industrial consumer at time t; is the spot market price; M is the number of photovoltaic units; L is the number of wind turbines; N is the number of energy storage units; is the nth wind turbine output at time t, Relc is the carbon benefit balancing value. 1.2 Analysis of the benefits of participating in the carbon market under the multi-coupling mechanism: The carbon market revenue calculation formula is: ; Wherein, is the total amount of carbon dioxide emissions offset by new energy power generation; , are the on-grid electricity of photovoltaic power generation and wettable power generation group used for the approved CCER, respectively; is the coal consumption rate of coal-fired power generation units; is the standard coal carbon emission coefficient; is the carbon dioxide emission coefficient per unit carbon combustion, and its value is 3.67; 1.3 Analysis of the benefits of participating in the green certificate market under the multi-coupling mechanism: The green certificate market revenue calculation formula is: ; wherein, is the green certificate market revenue; is the total feed-in power of approved photovoltaic green certificates; is the price of photovoltaic electricity green certificates; is the total feed-in power of approved wind power green certificates; is the total feed-in power of approved wind power green certificates; The revenue of the system participating in the carbon market under the multi-coupling mechanism is: ; wherein, is the income under the carbon market, is the CCER certificate transaction price; The total revenue of the industrial user and the power selling company participating in the "electricity-carbon-green certificate" multi-market is the sum of the revenues of the three markets, that is: 。 3. The method of claim 1, wherein: the method is an e-Carbon-Green certificate converged multi-market synergistic service equilibrium decision method. The specific steps of step 2 include: 2.1 Build the upper model: The upper model optimizes the charging and discharging state of the ES station in the system based on the "self-use priority" and "surplus power grid-connected" principles, optimizes the power declaration of the system participating in the electricity market, and improves the matching degree of grid-connected power and industrial load; ; The upper model needs to satisfy the constraints of unit output, energy storage output, energy storage state and energy storage capacity, etc.; The new energy unit output constraint is as follows: ; ; The load coverage constraint is: ; For the energy storage charging and discharging power constraint, the energy storage charging and discharging power must be less than the maximum charging and discharging power: ; ; wherein, , is the minimum charge and discharge power of the nth energy storage unit; , is the maximum charge and discharge power of the nth energy storage unit; , are the discharge and charge state variables of the nth ES unit, respectively; Due to the limitation of energy storage operation state, energy storage charging and discharging cannot be performed simultaneously: ; For the energy storage capacity constraint, the energy storage capacity must not exceed the upper and lower limits of the rated capacity; the capacity of the next moment is determined by the remaining capacity of the previous moment and the charging and discharging behavior at that moment; In order to ensure the parallelism of dispatching, the initial state and the final state of the capacity in a period are equal; ; ; ; wherein, Cn(t) is the capacity of the nth energy storage unit at time t; , respectively the charging and discharging efficiency of the energy storage; , respectively the capacity of the energy storage at the initial and final time. 2.2 Build the lower model: The objective function is to maximize the sum of carbon market and green certificate market revenue: ; The energy storage power station has energy loss in the charging and discharging process; GC and CCER verification is based on online power consumption, so the verified power consumption is the actual power generation minus the energy loss; the proportion of photovoltaic and wind power in energy storage, charging and discharging is determined based on the ratio of photovoltaic and wind turbine power generation at that time; ; ; ; wherein, is the percentage of photovoltaic unit output over time in t.

4. The method of claim 1, wherein the method is characterized by: The specific steps of step 3 include: 3.1 Data collection and standardization: collect the load data of the industrial user, the electricity market transaction data, the carbon market data and the green certificate market data for nearly 3 years; divide the training set and the test set according to different time scales and states, and standardize all input features; 3.2 Training of double-layer optimization model: (1) First, based on the collected data and policy requirements, set the initial value of the upper layer decision variable; (2) Take the initial decision of the upper layer as the hard constraint of the lower model, and solve the local optimal solution of the lower model, that is, the optimal operation strategy of the "follower" under the given rules of the "leader"; (3) Substitute the lower layer solution into the upper layer model, correct the decision variables of the upper layer, and realize the adaptive adjustment of the "leader" to the response of the "follower": repeat steps (2)-(3) until the following convergence conditions are met, and stop iteration.

5. The method of claim 1, wherein: the method is an e-Carbon-Green certificate converged multi-market synergistic service equilibrium decision method. The specific method of step 4 is: After training convergence, the effectiveness of the model needs to be verified through multi-dimensional verification, and the key parameters need to be calibrated; Feasibility verification: check whether the training results meet all the constraints. If there is a constraint violation, backtrack to adjust the initial value of the upper layer decision variable or the constraint weight; Optimality verification: compare the training results of different initial values to confirm that the final converged upper layer objective function value is globally optimal, avoiding optimization pitfalls caused by initial value deviation; Robustness verification: perturb the input data, retrain the model, and observe the change amplitude of the upper layer objective function value. Otherwise, the model constraint elasticity coefficient needs to be adjusted.

6. The method of claim 1, wherein the method is an e-Carbon-Green certificate integrated multi-market synergistic service equilibrium decision method. The step 5 model runs and outputs the daily optimal decision strategy, including: Allocation of electricity between long-term contracts and spot markets in the electricity market; CCER development and quota transaction recommendations in the carbon market; Green certificate purchase or sale strategy in the green certificate market to meet the renewable energy consumption responsibility RPS requirement.

Images