A park multi-producer and consumer low-carbon response transaction method and system considering an electricity-carbon market

Abstract

This invention relates to a low-carbon response trading method and system for multiple prosumers in industrial parks, considering the electricity-carbon market. The method includes: constructing a carbon emission flow model for the power grid and designing an electricity-carbon ratio model for energy storage devices; constructing a real-time collaborative decision-making model for electricity-carbon among multiple prosumers in the industrial park, aiming to minimize total operating costs; reconstructing the real-time collaborative decision-making model based on a Markov decision process and introducing event-driven criteria, reconstructing the global optimization objective function in the model as the sum of the operating cost in the current period and the optimal decision cost in subsequent stages; constructing and solving a cooperative game-theoretic real-time optimization trading decision-making model for multiple prosumers in the industrial park to obtain the real-time decision results of each prosumer; determining whether each prosumer meets the event-driven criteria; if so, executing the real-time decision result; otherwise, maintaining the decision result from the previous moment and waiting for the next round of optimization triggering. This invention can tap the proactive carbon reduction potential of prosumer groups and effectively address source-load uncertainties.

Description

Technical Field

[0001] This invention belongs to the field of power system technology, and in particular relates to a low-carbon response trading method and system for multi-producer-consumer transactions in industrial parks that considers the electricity-carbon market. Background Technology

[0002] As key participants in market transactions, industrial parks can promote urban energy transformation and energy conservation and emission reduction through multi-energy complementarity, production-consumption coordination, and demand response. Therefore, how to coordinate the low-carbon demand response trading decisions of multiple producers and consumers in the park, which have different time scales and need to be coupled, to achieve the optimal comprehensive benefits is a key challenge in realizing user-side electricity-carbon market decision-making synergy.

[0003] With the large-scale integration of distributed renewable energy and flexible loads, both the source and load sides exhibit strong stochastic fluctuations. This uncertainty not only exacerbates the dynamic changes in system state but also makes it difficult for deterministic decision-making methods to guarantee the timeliness and robustness of decisions. Furthermore, a comprehensive review of the carbon emission flow process in low-carbon dispatch is a crucial basis for evaluating the effectiveness of low-carbon dispatch. However, existing technologies all apply carbon emission flow models without considering the integration of energy storage devices. Due to the diversity and temporal coupling of energy storage device states, traditional carbon emission flow models struggle to accurately characterize the carbon emission characteristics of systems containing energy storage devices. This results in the inability to accurately trace carbon emission responsibility and hinders the full utilization of the flexibility of energy storage devices in trading decisions.

[0004] In model solving, distributed algorithms are a crucial optimization and control tool, particularly suitable for scenarios involving multi-energy system collaboration, multiple operating entities, information dispersion, or privacy protection requirements. With the large-scale grid connection of new energy sources and the increasing flexibility of load demand, stochastic factors in park systems are surging. However, current forecast information has high errors, posing a challenge to the practical applicability of day-ahead scheduling. Real-time scheduling, capable of dynamically adjusting operating strategies based on the latest status data, is of great significance for addressing uncertainty and volatility and improving system flexibility.

[0005] Real-time transaction decisions can dynamically optimize operational strategies based on the latest state information, playing a crucial role in enhancing the park system's responsiveness to uncertainty and volatility. However, in intraday real-time decision-making, the system state has a high dimension and contains multiple types of continuous variables. Conventional optimization algorithms are prone to the curse of dimensionality in high-dimensional state spaces. At the same time, each producer and consumer within the park has independent operational goals and data privacy requirements. While centralized decision-making improves efficiency, it also faces the problem of insufficient willingness to share information among stakeholders.

[0006] Therefore, it is urgent to construct a producer-consumer carbon measurement model based on dynamic electricity emission factors and a producer-consumer low-carbon demand response strategy based on an event-driven mechanism. This will enable accurate carbon emission characterization while achieving distributed offline training and real-time optimization, fully tapping the proactive carbon reduction potential of the producer-consumer group, and effectively addressing source-load uncertainties. Summary of the Invention

[0007] To address the challenges of accurately characterizing the carbon emission characteristics of systems containing energy storage devices using traditional carbon emission flow models, and the difficulty in efficiently solving real-time optimization decisions for multi-prosumer-consumer electricity-carbon collaboration in industrial parks that take into account the randomness of source loads, this invention provides an extended dynamic carbon metering model for multi-prosumer-consumers in industrial parks and a low-carbon demand response trading method that considers the electricity-carbon market.

[0008] The present invention adopts the following technical solution.

[0009] In a first aspect, this invention discloses a low-carbon response trading method for multi-producer-consumer transactions in industrial parks that considers the electricity-carbon market, the method comprising the following steps: Step 1: Construct a carbon emission flow model for the power grid and design an electricity-to-carbon ratio model for energy storage devices; Step 2: Based on the carbon emission flow model and the electricity-carbon ratio model, construct a real-time decision-making model for multi-producer-consumer electricity-carbon coordination in the industrial park with the goal of minimizing total operating costs; Step 3: Reconstruct the real-time decision-making model for multi-producer-consumer electricity-carbon collaboration in the park based on Markov decision process, and introduce event-driven criteria; Step 4: Decompose the global optimization objective function in the reconstructed real-time decision-making model of multi-producer-consumer electricity-carbon collaboration in the park into the sum of the operating cost of the current period and the optimal decision cost of the subsequent stage, and obtain the reconstructed objective function applicable to sequential decision-making. Step 5: Based on the reconstructed objective function, construct a real-time optimized transaction decision model for multi-prosumer cooperative game in the park; Step 6: Solve the real-time optimization decision model of the multi-prosumer cooperative game in the park to obtain the real-time decision results of each prosumer; determine whether each prosumer meets the event-driven criterion. If it does, execute the real-time decision result; otherwise, maintain the decision result of the previous moment and wait for the next round of optimization to be triggered.

[0010] More preferably, In step 1, the energy storage device's electricity-to-carbon ratio model is specifically as follows:

[0011]

[0012]

[0013] in, , These are energy storage devices e During the period t , t -1 is the carbon-to-electricity ratio of energy storage devices; , These are energy storage devices e During the period t -1、 t Available capacity; For energy storage devices e During the period t The amount of carbon emissions added; For energy storage devices e During the period t The amount of carbon emissions released; For energy storage devices e During the period t The charging power; For energy storage devices e The node in the time period t The node carbon emission intensity; For time intervals; For energy storage devices e During the period t The discharge power; For energy storage devices e When used as a power generation device during a certain period of time t Source-side carbon emission intensity; For energy storage devices e The discharge efficiency.

[0014] More preferably, In step 2, the total operating cost includes electricity purchase cost, wind and solar curtailment cost, load shedding cost, electricity market transaction cost, and carbon market transaction cost; the multi-producer-consumer electricity-carbon collaborative real-time decision-making model in the park has constraints including power balance constraints, wind and solar power output constraints, external grid power purchase and sale constraints, energy storage constraints, electricity trading volume constraints, carbon quota trading volume constraints, and carbon emission constraints. The carbon emission constraints are specifically as follows:

[0015] in, and The main body i During the period t and time period Carbon emissions; Carbon potential of purchased electricity as the main body i and main body jDuring the period t Carbon quota trading volume, I For the set of all subjects, as the main body i During the period t Electricity purchased from the power grid.

[0016] More preferably, In step 3, the real-time collaborative decision-making model for electricity and carbon emissions among multiple producers and consumers in the park, reconstructed based on the Markov decision process, specifically includes: using the energy storage capacity, carbon emissions, actual renewable energy output, and load power of each entity in the park as state variables, and the energy storage charging and discharging power, electricity trading volume, carbon quota trading volume, and actual renewable energy consumption as decision variables, combined with external information, to construct a state transition equation to characterize the dynamic operating characteristics of the system; the external information includes the prediction error and day-ahead forecast value of renewable energy, and the prediction error and day-ahead forecast value of load.

[0017] More preferably, In step 3, the event-driven criterion is specifically as follows:

[0018] in, as the main body i During the period t In terms of carbon emissions; as the main body i The carbon emission threshold.

[0019] More preferably, In step 4, the reconstructed objective function is specifically as follows:

[0020] in, From t The minimum total running cost function from the start of the trading session to the end of the transaction; , They are respectively t Decision variables and state variables for different time periods; for t Time period in state ,decision making Current operating costs for the current period; In order to be in t Make decisions during the time period The new state that the system then enters; Indicates a new state The minimum cumulative operating cost from start to finish of the transaction is the optimal decision cost for subsequent stages.

[0021] More preferably, The real-time optimization transaction decision model for multi-producer-consumer cooperative game in the park includes a cooperative benefit maximization sub-model and an energy transaction payment sub-model. The cooperative benefit maximization sub-model is used to determine the optimal electricity and carbon quota trading volume among the various entities, and the energy transaction payment sub-model is used to determine the optimal electricity and carbon quota trading price among the various entities. The two sub-models are solved in a distributed iterative manner using an augmented Lagrangian function.

[0022] More preferably, Step 6 involves solving the real-time optimization decision-making model for multi-prosumer cooperative game in the park, which includes two stages: day-ahead offline training and day-ahead real-time decision-making. Specifically: During the recent offline training phase: The PR-ADMM algorithm is used to solve the cooperation benefit maximization sub-model and the real-time energy transaction payment sub-model to obtain the optimal electricity and carbon quota trading volume and the optimal electricity and carbon quota trading price among the participants. In the reconstructed objective function, the high-dimensional state variables related to energy flow are aggregated into energy storage capacity, and the high-dimensional state variables related to carbon flow are aggregated into carbon emissions. Based on the approximation function of the convex piecewise linear function, the equivalent value of the optimal decision cost in the subsequent stage of the reconstructed objective function is modeled to obtain the reconstructed objective function after the equivalent value modeling. Based on the optimal electricity and carbon quota trading volume and the optimal electricity and carbon quota trading price among the aforementioned entities, calculate the slope of each segment in the reconstructed objective function after equivalent value modeling; During the intraday real-time decision-making phase: Based on the slopes of each segment calculated during the offline training phase, the reconstruction objective function after the equivalent value model is constructed and solved to obtain the real-time decision results of each producer and consumer.

[0023] More preferably, The reconstructed objective function after modeling the equivalent values ​​is specifically as follows:

[0024] in, D The total number of segments in a piecewise linear function; and Energy storage b and main body i During the period t No. d The slope of the segment; For energy storage b During the period t No. d Segmented energy storage capacity; as the main body i During the period t No.d Segmented carbon emissions; D The total number of segments in a piecewise linear function; B A collection of energy storage devices.

[0025] More preferably, The slope of each segment in the reconstructed objective function after the equivalent value modeling is determined as follows:

[0026]

[0027] in, k This represents the number of iterations. , The first k In this iteration, energy storage devices b ,main body i During the period and segmentation The slope of the slope; , The first k In the -1 iteration, energy storage devices b ,main body i During the period and segmentation Lower segment slope; , They were respectively in the second k Energy storage during the next iteration b ,main body i During the period t The sampled estimate of the slope of the lower piecewise linear function; Divide the linear function into segments; To update the step size, and satisfy the following conditions: .

[0028] More preferably, The sampled estimate of the slope of the piecewise linear function is calculated as follows: A perturbation is applied to the energy storage state, and the slope sampling estimate of each energy storage device in the piecewise linear function is calculated in two scenarios based on whether the perturbation affects the energy storage charging and discharging power. When the charging and discharging power is not affected, the slope sampling estimate of each energy storage device is calculated as follows:

[0029] in, For the first k In this iteration, energy storage devices b During the period t and segmentation The sampled estimate of the slope of the linear function under the given conditions; For the first k In the -1 iteration, energy storage devices b During the period t and segmentation d The slope of the slope; For the first k In this iteration, energy storage devices b During the period t and segmentation d The amount of energy stored below; For the first k In this iteration, energy storage devices b During the period t The energy storage capacity; When considering factors affecting charging and discharging power, the slope sampling estimate for each energy storage device is calculated as follows:

[0030] in, For energy storage b The marginal electricity price at the node where it is located; A perturbation is applied to the carbon emission status, and the slope sampling estimate of each subject in the piecewise linear function is calculated in two scenarios, depending on whether the perturbation affects the generation of carbon emissions within the park. When the generation of carbon emissions within the park is not affected, the slope sampling estimate for each entity is calculated as follows:

[0031] in, For the first k In the next iteration, the main body i During the period t Sampling estimate of the slope of a linear function; For the first k In the -1st iteration, the main body i During the period t and segmentation d The slope of the slope; as the main body i During the period t and segmentation d Carbon emissions; For the first k In the next iteration, the main body i During the period t Carbon emissions; When influencing carbon emissions within the park, the slope sampling estimate for each entity is calculated as follows:

[0032] in, For carbon-producing units g The marginal electricity price of the node.

[0033] Secondly, this invention discloses a low-carbon response trading system for multiple producers and consumers in industrial parks that considers the electricity-carbon market based on the aforementioned method. The system includes a power grid carbon emission flow model construction module, a real-time decision-making model construction module for multiple producers and consumers in industrial parks that coordinates electricity and carbon emissions, a real-time decision-making model reconstruction module for multiple producers and consumers in industrial parks that coordinates electricity and carbon emissions, an objective function reconstruction module, a real-time optimization trading decision-making model construction module for cooperative game theory among multiple producers and consumers in industrial parks, and a solution module for the real-time optimization trading decision-making model for cooperative game theory among multiple producers and consumers in industrial parks. A module for building a carbon emission flow model for power grids is used to construct a carbon emission flow model for power grids and design an energy storage device's carbon-to-electricity ratio model. The module for constructing a real-time decision-making model for multi-producer-consumer electricity-carbon coordination in industrial parks, based on the carbon emission flow model and the electricity-carbon ratio model, constructs a real-time decision-making model for multi-producer-consumer electricity-carbon coordination in industrial parks with the goal of minimizing total operating costs. The module for reconstructing the real-time decision-making model of multi-producer-consumer electricity-carbon collaboration in the industrial park reconstructs the model based on the Markov decision process and introduces event-driven criteria. The objective function reconstruction module decomposes the global optimization objective function in the real-time decision-making model of multi-producer-consumer electricity-carbon collaboration in the park into a recursive combination of the current operating cost and the optimal decision cost in the subsequent stage, and obtains a reconstructed objective function suitable for sequential decision-making. The module for constructing a real-time optimized transaction decision model for multi-prosumer cooperative game in the industrial park is based on the reconstructed objective function to construct a real-time optimized transaction decision model for multi-prosumer cooperative game in the industrial park. The module for solving the real-time optimization transaction decision model of the multi-prosumer cooperative game in the park solves the real-time optimization decision model of the multi-prosumer cooperative game in the park to obtain the real-time decision results of each prosumer; it determines whether each prosumer meets the event-driven criterion. If it does, the real-time decision result is executed; otherwise, the decision result of the previous moment is maintained and the next round of optimization is waited for.

[0034] Thirdly, the present invention provides a terminal, including a processor and a storage medium; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the method according to any one of the first aspects of the present invention.

[0035] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in any one of the first aspects of the present invention.

[0036] The beneficial effects of this invention are that, compared with the prior art, Based on carbon emission theory, user carbon emission traceability is achieved through power flow tracking. Power flow is further refined according to producer-consumer contracts. Dynamic power consumption carbon emission factors that reflect the spatiotemporal differences in electricity consumption carbon emissions are calculated. A power grid carbon emission flow model is constructed, clarifying the correspondence between "carbon flow" and "power flow," thus overcoming the shortcomings of the existing power grid average emission factor method. The concept of electricity-carbon ratio (ECR) for energy storage devices is proposed to characterize the carbon emission characteristics of energy storage devices. A unified carbon emission flow model for multiple producer-consumers in a park that takes into account energy storage devices is constructed, fully exploring the potential for coordination between supply and storage. The Markov decision process (MDP) reconstruction of the real-time optimization trading decision model for electricity-carbon collaboration among industrial park producers and consumers is realized. An event-driven mechanism-based response strategy for low-carbon energy use behavior of producers and consumers is proposed. Furthermore, a real-time decision model for multiple producers and consumers in industrial parks based on the Bellman optimality principle is constructed, which empowers producers and consumers to effectively perceive the spatiotemporal differences in their indirect carbon emission responsibilities. A real-time optimization decision-making model for cooperative game among multiple producers and consumers in a park is constructed. The PR-ADMM solution algorithm is introduced, and a distributed value function slope update method based on local information is proposed. This solves the problem of the model dimension surge and difficulty in efficient and accurate solution, realizes distributed offline training and real-time optimization, and effectively copes with the uncertainty of source load. Attached Figure Description

[0037] Figure 1 This is a flowchart illustrating the low-carbon response trading method for multi-producer-consumer in industrial parks that considers the electricity-carbon market in this invention. Figure 2 This is a schematic diagram illustrating the specific process of solving the real-time optimization decision-making model for multi-prosumer cooperative game in the industrial park in Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of the multi-producer-consumer architecture in the industrial park according to Embodiment 1 of the present invention; Figure 4 This is the real-time decision-making process for multi-producer-consumer electricity-carbon collaboration in the industrial park in Embodiment 1 of the present invention; Figure 5 This is a piecewise linear approximation process of the energy storage capacity value function in Embodiment 1 of the present invention; Figure 6 This is a diagram illustrating the piecewise linear approximation process of the carbon emission value function in Embodiment 1 of the present invention; Figure 7 This is a diagram showing the operation of energy storage in the park before and after the addition of disturbance in Embodiment 1 of the present invention; Figure 8 This is a graph showing the change in carbon emissions in the park before and after the addition of disturbance in Embodiment 1 of the present invention. Detailed Implementation

[0038] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.

[0039] like Figure 1 As shown, this invention discloses a low-carbon response trading method for multi-producer-consumers in industrial parks that considers the electricity-carbon market, comprising the following steps: Step 1: Construct a carbon emission flow model for the power grid and design an electricity-to-carbon ratio model for energy storage devices; The energy storage device's electricity-to-carbon ratio model is as follows:

[0040]

[0041]

[0042] in, , These are energy storage devices e During the period t , t -1 is the carbon-to-electricity ratio of energy storage devices; , These are energy storage devices e During the period t -1、 t Available capacity; For energy storage devices e During the period t The amount of carbon emissions added; For energy storage devices e During the period t The amount of carbon emissions released; For energy storage devices e During the period t The charging power; For energy storage devices e The node in the time period t The node carbon emission intensity; For time intervals; For energy storage devices e During the period t The discharge power; For energy storage devices e When used as a power generation device during a certain period of timet Source-side carbon emission intensity; For energy storage devices e The discharge efficiency.

[0043] Step 2: Based on the carbon emission flow model and the electricity-carbon ratio model, construct a real-time decision-making model for multi-producer-consumer electricity-carbon coordination in the industrial park with the goal of minimizing total operating costs; The total operating cost includes electricity purchase cost, wind and solar curtailment cost, load shedding cost, electricity market transaction cost, and carbon market transaction cost; the multi-producer-consumer electricity-carbon collaborative real-time decision-making model of the park has constraints including power balance constraints, wind and solar power output constraints, external grid power purchase and sale constraints, energy storage constraints, electricity trading volume constraints, carbon quota trading volume constraints, and carbon emission constraints. The carbon emission constraints are specifically as follows:

[0044] in, and The main body i During the period t and time period Carbon emissions; Carbon potential of purchased electricity as the main body i and main body j During the period t Carbon quota trading volume, I For the set of all subjects, as the main body i During the period t Electricity purchased from the power grid.

[0045] Step 3: Reconstruct the real-time decision-making model for multi-producer-consumer electricity-carbon collaboration in the park based on Markov decision process, and introduce event-driven criteria; The real-time collaborative decision-making model for electricity and carbon emissions among multiple producers and consumers in the industrial park, reconstructed based on the Markov decision process, specifically includes: using the energy storage capacity, carbon emissions, actual renewable energy output, and load power of each entity in the park as state variables, and the energy storage charging and discharging power, electricity trading volume, carbon quota trading volume, and actual renewable energy consumption as decision variables, combined with external information, to construct a state transition equation to characterize the dynamic operating characteristics of the system; the external information includes the prediction error and day-ahead forecast value of renewable energy, and the prediction error and day-ahead forecast value of load.

[0046] The event-driven criteria are as follows:

[0047] in, as the main body i During the periodt In terms of carbon emissions; as the main body i The carbon emission threshold.

[0048] Step 4: Decompose the global optimization objective function in the reconstructed real-time decision-making model of multi-producer-consumer electricity-carbon collaboration in the park into the sum of the operating cost of the current period and the optimal decision cost of the subsequent stage, and obtain the reconstructed objective function applicable to sequential decision-making. The reconstruction objective function is specifically as follows:

[0049] in, From t The minimum total running cost function from the start of the trading session to the end of the transaction; , They are respectively t Decision variables and state variables for different time periods; for t Time period in state ,decision making Current operating costs for the current period; In order to be in t Make decisions during the time period The new state that the system then enters; Indicates a new state The minimum cumulative operating cost from start to finish of the transaction is the optimal decision cost for subsequent stages.

[0050] Step 5: Based on the reconstructed objective function, construct a real-time optimized transaction decision model for multi-prosumer cooperative game in the park; The real-time optimization transaction decision model for multi-producer-consumer cooperative game in the park includes a cooperative benefit maximization sub-model and an energy transaction payment sub-model. The cooperative benefit maximization sub-model is used to determine the optimal electricity and carbon quota trading volume among the various entities, and the energy transaction payment sub-model is used to determine the optimal electricity and carbon quota trading price among the various entities. The two sub-models are solved in a distributed iterative manner using an augmented Lagrangian function.

[0051] Step 6: Solve the real-time optimization decision model of the multi-prosumer cooperative game in the park to obtain the real-time decision results of each prosumer; determine whether each prosumer meets the event-driven criterion. If it does, execute the real-time decision result; otherwise, maintain the decision result of the previous moment and wait for the next round of optimization to be triggered.

[0052] The real-time optimization decision-making model of multi-prosumer cooperative game in the aforementioned park is solved, including two stages: day-ahead offline training and intraday real-time decision-making, specifically as follows: During the recent offline training phase: The PR-ADMM algorithm is used to solve the cooperation benefit maximization sub-model and the real-time energy transaction payment sub-model to obtain the optimal electricity and carbon quota trading volume and the optimal electricity and carbon quota trading price among the participants. In the reconstructed objective function, the high-dimensional state variables related to energy flow are aggregated into energy storage capacity, and the high-dimensional state variables related to carbon flow are aggregated into carbon emissions. Based on the approximation function of the convex piecewise linear function, the equivalent value of the optimal decision cost in the subsequent stage of the reconstructed objective function is modeled to obtain the reconstructed objective function after the equivalent value modeling. Based on the optimal electricity and carbon quota trading volume and the optimal electricity and carbon quota trading price among the aforementioned entities, calculate the slope of each segment in the reconstructed objective function after equivalent value modeling; During the intraday real-time decision-making phase: Based on the slopes of each segment calculated during the offline training phase, the reconstruction objective function after the equivalent value model is constructed and solved to obtain the real-time decision results of each producer and consumer.

[0053] Furthermore, the reconstruction objective function after modeling the equivalent values ​​is specifically as follows:

[0054] in, D The total number of segments in a piecewise linear function; and Energy storage b and main body i During the period t No. d The slope of the segment; For energy storage b During the period t No. d Segmented energy storage capacity; as the main body i During the period t No. d Segmented carbon emissions; D The total number of segments in a piecewise linear function; B A collection of energy storage devices.

[0055] The slope of each segment in the reconstructed objective function after the equivalent value modeling is determined as follows:

[0056]

[0057] in, k This represents the number of iterations. , The first k In this iteration, energy storage devicesb ,main body i During the period and segmentation The slope of the slope; , The first k In the -1 iteration, energy storage devices b ,main body i During the period and segmentation Lower segment slope; , They were respectively in the second k Energy storage during the next iteration b ,main body i During the period t The sampled estimate of the slope of the lower piecewise linear function; Divide the linear function into segments; To update the step size, and satisfy the following conditions: .

[0058] Furthermore, The sampled estimate of the slope of the piecewise linear function is calculated as follows: A perturbation is applied to the energy storage state, and the slope sampling estimate of each energy storage device in the piecewise linear function is calculated in two scenarios based on whether the perturbation affects the energy storage charging and discharging power. When the charging and discharging power is not affected, the slope sampling estimate of each energy storage device is calculated as follows:

[0059] in, For the first k In this iteration, energy storage devices b During the period t and segmentation The sampled estimate of the slope of the linear function under the given conditions; For the first k In the -1 iteration, energy storage devices b During the period t and segmentation d The slope of the slope; For the first k In this iteration, energy storage devices b During the period t and segmentation d The amount of energy stored below; For the first k In this iteration, energy storage devices b During the period t The energy storage capacity; When considering factors affecting charging and discharging power, the slope sampling estimate for each energy storage device is calculated as follows:

[0060] in, For energy storage b The marginal electricity price at the node where it is located; A perturbation is applied to the carbon emission status, and the slope sampling estimate of each subject in the piecewise linear function is calculated in two scenarios, depending on whether the perturbation affects the generation of carbon emissions within the park. When the generation of carbon emissions within the park is not affected, the slope sampling estimate for each entity is calculated as follows:

[0061] in, For the first k In the next iteration, the main body i During the period t Sampling estimate of the slope of a linear function; For the first k In the -1st iteration, the main body i During the period t and segmentation d The slope of the slope; as the main body i During the period t and segmentation d Carbon emissions; For the first k In the next iteration, the main body i During the period t Carbon emissions; When influencing carbon emissions within the park, the slope sampling estimate for each entity is calculated as follows:

[0062] in, For carbon-producing units g The marginal electricity price of the node.

[0063] Based on the optimal electricity and carbon quota trading volume and the optimal electricity and carbon quota trading price among the aforementioned entities, the slopes of each segment in the minimum total operating cost function expression after equivalent value modeling are calculated.

[0064] This invention also discloses a low-carbon response trading system for multiple producers and consumers in industrial parks that considers the electricity-carbon market based on the aforementioned method. The system includes a power grid carbon emission flow model construction module, a real-time decision-making model construction module for multiple producers and consumers in industrial parks that coordinates electricity and carbon emissions, a real-time decision-making model reconstruction module for multiple producers and consumers in industrial parks that coordinates electricity and carbon emissions, an objective function reconstruction module, a real-time optimization trading decision-making model construction module for cooperative game theory among multiple producers and consumers in industrial parks, and a solution module for the real-time optimization trading decision-making model for cooperative game theory among multiple producers and consumers in industrial parks. A module for building a carbon emission flow model for power grids is used to construct a carbon emission flow model for power grids and design an energy storage device's carbon-to-electricity ratio model. The module for constructing a real-time decision-making model for multi-producer-consumer electricity-carbon coordination in industrial parks, based on the carbon emission flow model and the electricity-carbon ratio model, constructs a real-time decision-making model for multi-producer-consumer electricity-carbon coordination in industrial parks with the goal of minimizing total operating costs. The module for reconstructing the real-time decision-making model of multi-producer-consumer electricity-carbon collaboration in the industrial park reconstructs the model based on the Markov decision process and introduces event-driven criteria. The objective function reconstruction module decomposes the global optimization objective function in the real-time decision-making model of multi-producer-consumer electricity-carbon collaboration in the park into a recursive combination of the current operating cost and the optimal decision cost in the subsequent stage, and obtains a reconstructed objective function suitable for sequential decision-making. The module for constructing a real-time optimized transaction decision model for multi-prosumer cooperative game in the industrial park is based on the reconstructed objective function to construct a real-time optimized transaction decision model for multi-prosumer cooperative game in the industrial park. The module for solving the real-time optimization transaction decision model of the multi-prosumer cooperative game in the park solves the real-time optimization decision model of the multi-prosumer cooperative game in the park to obtain the real-time decision results of each prosumer; it determines whether each prosumer meets the event-driven criterion. If it does, the real-time decision result is executed; otherwise, the decision result of the previous moment is maintained and the next round of optimization is waited for.

[0065] Example 1: like Figure 1 As shown, this invention discloses a low-carbon response trading method for multi-producer-consumers in industrial parks that considers the electricity-carbon market. The method includes the following steps: Step 1: Construct a carbon emission flow model for the power grid and design an electricity-to-carbon ratio model for energy storage devices; specifically including: Step 1.1: Based on the carbon emission flow theory, construct a refined carbon emission flow model for the power grid and analyze the correspondence between "carbon flow" and "power flow". The amount of carbon emissions passing through a network branch or node per unit time (in tCO2 / h) can be characterized by the carbon emission flow rate (CEFR), specifically:

[0066] in, Carbon emission flow rate; F Carbon emissions flowing through network branches or nodes; t For time.

[0067] Carbon intensity (CI) is used to characterize the amount of carbon emissions per unit of energy (unit: tCO2 / (MW)). Based on its characteristics, carbon emission intensity (CCI) can be divided into generation carbon intensity (GCI), which characterizes the carbon emissions associated with each unit of energy generated at the source; branch carbon intensity (BCI), which characterizes the carbon emissions associated with each unit of energy flowing along a branch; and node carbon intensity (NCI), which characterizes the average carbon emissions associated with each unit of energy injected into a node. Furthermore, to characterize the impact of producer-consumer contracts, GCI is divided into contract generation carbon intensity (CGCI) and nature generation carbon intensity (NGCI). CGCI characterizes the carbon emissions associated with each unit of energy from contracted power sources within producer-consumer contracts; NGCI characterizes the carbon emissions associated with each unit of energy from all power sources not participating in power contracts outside of producer-consumer contracts.

[0068] node i During the period t The NCI is shown below:

[0069] In the formula, , , and Power system nodes i Branch roads Contract power units cg and other non-contract power sources ng During the period t carbon emission intensity; For nodes i Branches with the first node ; For nodes i Connected contracted units cg ; For nodes i Connected uncontracted units ng ; For time period t Flowing through the branch road The electrical power; For time period t Aircraft crews participating in the contract cg The power; For time period t Aircraft crews not participating in the contract ng The power.

[0070] Step 1.2: Design the ECR model for energy storage equipment, propose the concept of ECR ​​for energy storage equipment, characterize the relationship between the amount of electricity in the energy storage equipment and the amount of carbon emissions absorbed, and accurately depict the carbon emission situation of multiple producers and consumers in the park; Energy storage devices have both charging and discharging states. When charging, they act as a load, absorbing some carbon emissions; when discharging, they act as a power source, releasing some carbon emissions. Therefore, this invention, referencing the state of charge (SOC) of energy storage devices, proposes the concept of the carbon-to-electricity ratio of energy storage devices to characterize the relationship between the amount of electricity stored and the amount of carbon emissions absorbed.

[0071] When an energy storage device is charging, carbon emissions are generated along with the electrical energy supplied to the device.

[0072] In the formula, For energy storage devices e During the period t The amount of carbon emissions added; For energy storage devices e During the period t The charging power; For energy storage devices e The node in the time period t Node carbon emission intensity (NCI); For time intervals.

[0073] When an energy storage device is in a discharging state, carbon emissions are released along with electrical energy from the energy storage device:

[0074] In the formula, For energy storage devices e During the period t The amount of carbon emissions released; For energy storage devices e During the period t The discharge power; For energy storage devices e When used as a power generation device during a certain period of time t Source-side carbon emission intensity (GCI); For energy storage devices e The discharge efficiency; For energy storage devices e During the period t -1 ECR (Electrocarbon Ratio).

[0075] Specifically, the electrocarbon ratio (ECR) of an energy storage device is defined as:

[0076] In the formula, , These are energy storage devices e During the period t , t -1 available capacity (in MW) h).

[0077] Step 2: Based on the carbon emission flow model and the electricity-carbon ratio model, construct a real-time decision-making model for multi-producer-consumer electricity-carbon collaboration in the industrial park with the goal of minimizing total operating costs.

[0078] The park aims to minimize total operating costs, which include electricity purchase costs, wind and solar curtailment costs, load shedding costs, electricity market transaction costs, and carbon market transaction costs.

[0079] The real-time decision-making model for multi-producer-consumer electricity-carbon collaboration in the industrial park can be specifically represented as follows:

[0080]

[0081]

[0082] In the formula, I It is a collection of various producers and consumers; T This represents the total number of time periods scheduled. V i and W i The main body i A combination of photovoltaic power sources and distributed wind power; as the main body i During the period t Total operating costs; , , ,and The main body i During the period t The costs of electricity purchase, wind and solar curtailment, load shedding, and related to the main body j The transaction costs, including electricity market transaction costs and carbon market transaction costs; The electricity purchase price for the upper-level energy grid; as the main body i During the period t Electricity purchased from the power grid; and These are the penalty cost coefficients for abandoning solar and wind power; and The main body i Photovoltaic units v Wind turbine w During the period t The actual power generation capacity; and The main body i Photovoltaic units v Wind turbine w During the period t The actual absorption capacity; This is the penalty cost factor for load shedding; as the main body i During the period t The amount of load shedding; and The main body i During the period t With the main body j Trading volume of electricity and carbon allowances; and The main body i During the period t With the main body j The trading prices corresponding to traded electricity and carbon allowances.

[0083] The real-time decision-making model for multi-producer-consumer electricity-carbon collaboration in the industrial park includes the following constraints: power balance constraints, wind and solar power output constraints, external grid power purchase and sale constraints, energy storage constraints, electricity trading volume constraints, carbon quota trading volume constraints, and carbon emission constraints; specifically: Power balance constraints:

[0084] Wind and solar power output constraints: ,

[0085] External power purchase and sale constraints:

[0086] Energy storage constraints:

[0087] Constraints on electricity trading volume:

[0088] Carbon quota trading volume constraints:

[0089] Carbon emission constraints:

[0090] in, B i For energy storage devicesb A set; L i For electrical load l A set; and The main body i Energy storage devices b During the period t The charging and discharging power; as the main body i electrical load l During the period t The load; This is the upper limit for purchased electricity; and These are energy storage devices b Upper limits for charging and discharging power; and The main body i Energy storage devices b During the period t and time period The amount of electricity; and These are energy storage devices b Maximum and minimum battery level; and The main body i Energy storage devices b Electricity consumption at the end and beginning of the trading cycle; and These are the upper and lower limits for electricity trading volume, respectively. and These are the upper and lower limits for carbon quota trading volume; , The main body i During the period t With the main body j Trading volume of electricity and carbon allowances; and The main body i During the period t and time period Carbon emissions; Carbon potential for purchased electricity; as the main body i During the period t Electricity purchased from the power grid.

[0091] Step 3: Based on the Markov decision process, reconstruct the real-time decision-making model of multi-producer-consumer electricity-carbon collaboration in the park, introduce event-driven criteria, and propose a response strategy for low-carbon energy use behavior of producers and consumers based on the event-driven mechanism.

[0092] The multi-prosumer electricity-carbon collaborative real-time decision-making model constructed in this invention is an optimization model with multi-time-period coupling and correlation. Wind power, photovoltaic output, and load demand in the industrial park exhibit strong randomness. Since current forecasting technologies cannot accurately predict intraday actual values, prosumer decision-makers can only make decisions based on the system's current state information. The multi-prosumer architecture in the industrial park is as follows: Figure 3 As shown. Therefore, the real-time decision-making process of each producer and consumer is a multi-stage sequential decision-making process, which can be reconstructed into an MDP.

[0093] Based on the reconstructed Markov decision process, the park's multi-producer-consumer electricity-carbon collaborative real-time decision model uses the energy storage capacity, carbon emissions, actual renewable energy output, and load power of each entity in the park as state variables, and the energy storage charging and discharging power, electricity trading volume, carbon quota trading volume, and actual renewable energy consumption as decision variables. Combined with external information, a state transition equation is constructed to characterize the dynamic operating characteristics of the system. The external information includes the prediction error and day-ahead forecast value of renewable energy, and the prediction error and day-ahead forecast value of load. The renewable energy includes wind turbine output and photovoltaic output.

[0094] Specifically, the real-time decision-making process (MDP) reconstruction of each prosumer in the multi-prosumer-consumer electricity-carbon collaborative real-time decision-making model includes four key parts, namely state variables. Decision variables Including prediction error vector vector of predicted values External information factors and state transition equations characterizing state change processes Specifically:

[0095]

[0096]

[0097]

[0098]

[0099] In the formula, , , , as well as Each main energy storage device is located during a specific time period. t The amount of electricity, and the time period of each entity t Carbon emissions of each main wind turbine unit during the time period t Actual power generation, and the time period of each main photovoltaic unit t Actual power generation and each main time periodt The electrical load; , Each main energy storage device is located during a specific time period. t The charging and discharging power; , Each subject in the time period t Electricity purchased from the power grid, and the time periods of various entities t The trading volume of carbon allowances; , Each main wind turbine unit is located during a specific time period. t Actual power absorption capacity and photovoltaic units during the time period t The actual absorption capacity; , and These represent wind power output, solar power output, and electrical load during different time periods. t The prediction error; , and These represent wind power output, solar power output, and electrical load during different time periods. t The current forecast value; For decision-making time steps, i For vector index; , Time periods t With the next period The first of the state variables i One component; , Time periods t The first decision variable i +1, No. i +2 components; For time period t The first of the predicted values ​​vectors of the previous day among the external information factors i -2 components; For the next period The first element in the prediction error vector among external information factors i -2 components; Carbon potential for purchased electricity.

[0100] Furthermore, to enable prosumers to effectively perceive the spatiotemporal differences in their indirect carbon emission responsibilities and fully explore the proactive carbon reduction potential of the prosumer group, an event-driven criterion is introduced. A prosumer low-carbon energy consumption behavior response strategy based on an event-driven mechanism is proposed. When carbon emissions exceed a preset threshold, the event trigger condition is activated, driving the prosumer to transition from the current state to the next state. If carbon emissions do not exceed the preset threshold, the prosumer does not take any optimization decisions at that moment. This ensures that prosumers can rationally formulate their response behavior according to their goals and generate certain benefits or costs during the transition process, ultimately achieving the optimal action combination across all states. The event-driven judgment condition is as follows:

[0101] in, as the main body i During the period t In terms of carbon emissions; as the main body i The carbon emission threshold; in this invention, the subject refers to the prosumers within the park, and each prosumer corresponds to an independent decision-making subject; the subject i The carbon emission threshold is determined as follows: First, each producer and consumer must ensure that during the cycle... Total carbon emissions The threshold is not to exceed the initial total carbon allowance; therefore, this threshold is based on the initial total carbon allowance for producers and consumers, obtained by averaging it across multiple time periods to obtain the average permissible carbon emissions; then, a safety factor is applied. Multiplying this value amplifies it as a trigger threshold; this method ensures inherent consistency between event triggering and carbon quota management. ,in, The value is determined based on the relative fluctuation of carbon emissions. In this embodiment, the preferred range is between 1.05 and 1.3.

[0102] More preferably, the load characteristics and carbon emissions of each prosumer were considered. Due to differences in load type and energy consumption structure, carbon emissions exhibit significant differences among prosumers. During the offline training phase, statistical analysis was performed on 500 Monte Carlo scenarios to determine the historical fluctuation range and standard deviation of carbon emissions for each prosumer. This data was used to establish differentiated threshold parameters for different prosumers. Specifically, prosumers with high carbon emissions used relatively high thresholds to prevent frequent triggering, while prosumers with low carbon emissions used lower thresholds to maintain sufficient response sensitivity.

[0103] The real-time decision-making process for multi-product-consumer electricity-carbon collaboration in industrial parks based on MDP is as follows: Figure 4 As shown, the specific decision-making process is as follows: Step 3.1: Prosumers in the park, based on... t Status of the time period Make the optimal decision ; Step 3.2: Prosumers in the park combine known external information factors and It then makes a judgment based on the event-driven criteria. If the requirements are met, it transitions to the next state according to the state transition equation. Therefore, consumers make optimization decisions at any given moment; otherwise, they do not make optimization decisions. Step 3.3: Following the order of the beginning and end of each time period, make decisions sequentially to obtain the real-time decision results of multiple producers and consumers in the park within the decision-making cycle.

[0104] Step 4: Decompose the global optimization objective function in the multi-producer-consumer electricity-carbon collaborative real-time decision-making model into the sum of the operating cost of the current period and the optimal decision cost of the subsequent stage, and obtain the reconstructed objective function applicable to sequential decision-making; Specifically, to obtain the globally optimal decision, the optimal decision for the current time period is sequentially solved based on the Bellman optimality principle to achieve the optimal real-time decision for the entire trading cycle, and the objective function is reconstructed:

[0105]

[0106] in, for t The period ends at the end of the trading time. T Minimum total operating cost for the time period; , They are respectively Decision variables and state variables for different time periods; for Time period in state ,decision making Current operating costs for the current period; In order to be in t Make decisions during the time period After that, the new state that the system enters is... , This represents the time interval between two adjacent decision moments; , They are respectively t Decision variables and state variables for different time periods; for t Time period in state ,decision making Current time period operating cost; value function Indicates a new state The minimum cumulative operating cost from start to finish of the transaction; In the state The optimal decision is the one that minimizes the total cost.

[0107] Step 5: Based on the reconstructed objective function, construct a real-time optimized trading decision model for multi-producer-consumer cooperative game in the park to determine the trading volume and price of electricity and carbon quotas for each time period; Specifically, a real-time optimization transaction decision model based on Nash bargaining theory is constructed for multi-prosumer cooperative game in the industrial park.

[0108] A real-time optimization trading decision model based on Nash bargaining theory is established for multi-prosumer cooperative game theory in industrial parks. This model determines the trading price and volume of electricity and carbon allowances in real time from the perspectives of overall rationality and individual rationality. Nash bargaining theory falls under the category of cooperative game theory and satisfies incentive compatibility conditions. Its standard model can be expressed as:

[0109] in, as the main body i During the period t Operating costs before participating in cooperative game theory; as the main body i During the period t The operating costs after participating in cooperative game theory, i.e. the point at which negotiations in cooperative game theory break down. Representing the subject i During the period t The cost reduction gained after participating in cooperative game theory.

[0110] The real-time optimization transaction decision model of the multi-prosumer cooperative game in the park is a nonlinear optimization problem. To achieve efficient distributed solution, it is decomposed into a cooperative benefit maximization sub-model and an energy transaction payment sub-model. The cooperative benefit maximization sub-model is used to determine the optimal electricity and carbon quota trading volume among the participants, and the energy transaction payment sub-model is used to determine the optimal electricity and carbon quota trading price among the participants. The two sub-models are solved in a distributed iterative manner using an augmented Lagrangian function.

[0111] To achieve distributed solution, the above sub-model is transformed into a corresponding sub-problem, and its augmented Lagrangian function is constructed to support iterative optimization.

[0112] For the subject i The augmented Lagrangian function of the real-time cooperation benefit maximization subproblem is:

[0113] In the formula, as the main body i With the main body j During the period tA generalized expression for trading electrical energy; as the main body i With the main body j During the period t A generalized expression of carbon trading quotas; Energy coupling constraints in time period t The Lagrange multipliers, the energy coupling constraints include constraints on electricity trading volume, carbon quota trading volume, carbon emission volume, and individual rationality constraints in Nash bargaining theory; To maximize the benefits of cooperation, the sub-problem is in the time period t The penalty parameter; as the main body i During the period t The remaining cost does not include the cost of electricity carbon trading; J It is the set of all subjects.

[0114] For the subject i The augmented Lagrangian function of the real-time energy transaction payment subproblem is:

[0115] In the formula, as the main body i With the main body j During the period t A generalized expression of electricity trading prices; as the main body i With the main body j During the period t A generalized expression of carbon quota trading prices; Sub-problems of energy trading payments during the period t Lagrange multipliers; Sub-problems of energy trading payments during the period t The penalty parameter; To maximize the benefits of cooperation, the sub-problem is in the time period t The optimal solution obtained; as the main body i exist t The optimal real-time trading solution is obtained based on the current trading situation of electricity and carbon quotas within the time period; the initial real-time trading prices of electricity and carbon quotas are both set to 0.

[0116] Step 6: Solve the real-time optimization decision model of the multi-prosumer cooperative game in the park to obtain the real-time decision results of each prosumer; determine whether each prosumer meets the event-driven criterion. If it does, execute the real-time decision result; otherwise, maintain the decision result of the previous moment and wait for the next round of optimization to be triggered.

[0117] A real-time transaction decision-making method for multiple prosumers in a park based on PR-ADMM-ADP is proposed. Combined with a distributed value function slope update method based on local information, the real-time optimization decision-making model of cooperative game among multiple prosumers in the park is solved to obtain the real-time decision results of each prosumer. Specifically, the solution process of this invention is divided into two parts: distributed collaborative optimization based on PR-ADMM and approximate iteration of the value function based on ADP. The slope of the value function is updated by combining local information, and the globally optimal real-time transaction solution is obtained quickly while protecting the privacy information of each producer and consumer.

[0118] Step 6.1: Distributed collaborative optimization based on PR-ADMM; The PR-ADMM algorithm is based on the PR splitting method, introducing 1 / 2 power multipliers on top of the classic ADMM. Essentially, it refers to Lagrange multipliers. Two iterations were performed, making the multiplier update more "aggressive," which typically leads to faster convergence. The specific update process is as follows:

[0119] In the formula, The number of iterations for the improved PR-ADMM; , The first , No. The main body during the next iteration i During the period t With the main body j A generalized expression for trading electrical energy; , For the first , No. The main body during the next iteration i During the period t With the main body j A generalized expression for trading carbon allowances; , , The first , No. , No. During the next iteration, the energy coupling constraint is in the time period t Lagrange multipliers; For the first In the next iteration, the subproblem of maximizing cooperation benefits occurs during the time period. t The penalty parameter; It is a relaxation factor, usually set in the range of (0, 1) to ensure convergence.

[0120] The PR-ADMM algorithm determines whether the iteration has ended by checking whether the original residual and the dual residual have reached convergence accuracy.

[0121] In the formula, For the first The next iteration obtains the time period t The original residuals; For the first The next iteration obtains the time period t The dual residual; and These represent the convergence accuracy of the original residual and the dual residual, respectively. In this embodiment, it is preferably 10. -4 , For the first In the next iteration, the subproblem of maximizing cooperation benefits occurs during the time period. t The penalty parameter.

[0122] The PR-ADMM algorithm is used to solve the subproblem of maximizing cooperation benefits in the real-time optimization decision-making model of multi-prosumer cooperative game in the park. If the calculated original residual or dual residual does not meet the convergence accuracy requirement, the multipliers are updated and the optimization calculation is repeated until both the original residual and the dual residual meet the convergence accuracy requirement. The calculated transaction volume of electricity and carbon quotas among the various entities is used as the initial value of the energy transaction payment subproblem, and the PR-ADMM algorithm is used to solve the energy transaction payment subproblem. If the calculated original residual or dual residual does not meet the convergence accuracy requirement, the multipliers are updated and the optimization calculation is repeated until both the original residual and the dual residual meet the convergence accuracy requirement. The optimal transaction price of electricity and carbon quota is then output.

[0123] Step 6.2: Construct an approximate value function of the optimal decision cost for the subsequent stages of the system based on ADP, and iteratively update the slope of the approximate value function using a distributed value function slope update method based on local information; For the reconstructed objective function in step 4, its state variables are precisely constructed. AND-value function The mapping relationship is crucial to ensuring optimal decision results. However, the large number and high dimensionality of state variables in the system often lead to a significant increase in computational burden. Therefore, state variable aggregation is used to alleviate the conflict between optimization accuracy and computational burden during model solving. In the multi-producer-consumer electricity-carbon collaborative real-time decision-making model, the stored energy in the energy flow... E b Carbon emissions in carbon streams These are all state variables with time-related characteristics, capable of extending the effect of a single time period to the entire trading cycle. High-dimensional state variables related to energy flow are aggregated into stored energy. E b High-dimensional state variables related to carbon flow are aggregated into carbon emissions. This allows us to characterize the impact of the optimal decision in the current state on future operations in a stochastic environment, thereby assisting in achieving the optimal real-time decision throughout the entire transaction cycle.

[0124] The process of aggregating high-dimensional state variables related to energy flow into stored energy capacity is described. E b High-dimensional state variables related to carbon flow are aggregated into carbon emissions. Specifically: The high-dimensional state variables refer to the variables related to energy flow and carbon flow. Specifically, high-dimensional state variables in energy flow include the State of Charge (SOC) of each entity and energy storage device, the charging and discharging power of the previous period, the voltage of each node, the power flow state, the output of each power source, and the load state. High-dimensional state variables in carbon flow include the carbon emission intensity and carbon emission rate of each node, the real-time carbon emissions of each generating unit, and the carbon emissions of each energy storage unit. The aggregation rule is as follows: For the energy flow component, the energy storage capacity is used as the only state variable with cross-period memory effect, and the high-dimensional state variables related to the energy flow are aggregated into the energy storage capacity. For the carbon flow component, the cumulative carbon emissions within the trading cycle are used as the core, and the carbon emission flow states of multiple nodes and branches are equivalently aggregated into the cumulative carbon emissions of the main component. This aggregation rule is established based on the energy balance relationship and the carbon emission accumulation relationship of energy storage, and can significantly reduce the state dimension while maintaining a complete representation of the impact on subsequent decisions.

[0125] The objective function of the real-time decision-making model for multi-producer-consumer electricity-carbon collaboration in industrial parks, based on the Bellman optimality principle, can be further expressed as:

[0126] in, For the time period t With energy storage b The relevant post-decision approximation function; For the time period t With the main body i Approximate function after decision-making related to carbon emissions. In order to be in t Time-of-use energy storage b The amount of electricity; In order to be in t main time period i Carbon emissions; B A collection of energy storage devices.

[0127] Specifically, the objective function of the real-time optimization decision-making model for multi-prosumer cooperative game in industrial parks based on PR-ADMM is a convex quadratic function, and all constraints are linear. To further reduce the computational burden in the value function fitting process, a convex piecewise linear function is used to approximate the value function, thereby reducing training time and improving decision results. The process of approximating the decoupled value function based on a convex piecewise linear function is as follows: Figure 5 and Figure 6 As shown, the total operating cost of the system can be... The equivalent value is:

[0128] in, D The total number of segments in a piecewise linear function; and Energy storage b and main body i During the period t No. d The slope of the segment; For energy storage b During the period t No. d Segmented energy storage capacity; as the main body i During the period t No. d Segmented carbon emissions.

[0129] No. d The segmented state of charge and carbon emissions satisfy the upper and lower limits and the total emission limits, respectively:

[0130] in, and The main body i The upper and lower limits of carbon emissions; and These are energy storage devices b Maximum and minimum battery level.

[0131] Figure 5 and Figure 6 The piecewise linear approximation processes of the value functions for the stored energy in the energy flow and the main carbon emissions in the carbon flow are presented respectively. (The main...) i and energy storage b For example, when energy storage b The decision to charge the vehicle is based on the state. Transition to state and through its approximation function This indicates the impact of the current action on the target decision for the remaining time period. (In energy storage) bWhile charging, based on the mapping relationship between energy flow and carbon flow, the main body... i Carbon emissions have also changed accordingly, with carbon emissions changing from state Transform into a state The remaining free carbon allowances available during the trading cycle are Approximate value function This reflects the impact of the remaining free carbon allowances on target decisions for the remaining period. Energy storage capacity. E Approximate function and main carbon emissions The approximate function together constitutes the overall approximate function of the main trading decision problem, which can satisfy the economic efficiency of the energy supply process while taking into account the carbon quota usage throughout the entire trading cycle, thus satisfying the low-carbon nature of the decision.

[0132] The value function is crucial for ensuring a globally optimal solution. It reflects the impact of the current state and decisions on the target decision for the remaining time period of the trading cycle. Therefore, it requires multiple offline training sessions to update the slope of the value function period by period. By embedding uncertainties into the value function through offline training to assist decision-making, the randomness of the source load in the system can be effectively addressed, ensuring that near-globally optimal decisions are made during real-time optimization.

[0133] Since the piecewise function is convex, value functions corresponding to the stored energy and the main carbon emissions are used to respectively... and The differential is used to calculate the sampled estimate of the slope, and then updated based on the slope value obtained in the previous iteration.

[0134]

[0135]

[0136]

[0137]

[0138] In the formula, k This represents the number of iterations. and They were respectively in the second k Energy storage during the next iteration b and main body i During the period t Sampling estimates of the slope of a piecewise linear function; Divide the linear function into segments; To update the step size, the value range is (0, 1]. For the first k In the next iteration, in the state Make a decision The immediate costs incurred , The first k In this iteration, energy storage devices b ,main body i During the period t The total optimal operating cost; , The first k In the -1 iteration, energy storage devices b ,main body i During the period t and segmentation d The slope of the slope; For the first k In this iteration, energy storage devices b During the period t and segmentation d The amount of energy stored below; as the main body i During the period t and segmentation d Carbon emissions; For the first k In this iteration, energy storage devices b During the period t The energy storage capacity; For the first k In the next iteration, the main body i During the period t Carbon emissions; , The first k In this iteration, energy storage devices b ,main body i During the period and segmentation The slope of the slope; , The first k In the -1 iteration, energy storage devices b ,main body i During the period and segmentation The slope of the lower segment.

[0139] More preferably, to avoid potential constraint conflicts when the stored energy simultaneously meets the upper and lower limit reconstruction constraints and the initial and final balance constraints, the constraint-aware value function estimation (CAVE) algorithm is further used to update the slope.

[0140] To integrate the value function update process with the proposed distributed real-time optimization trading decision framework for multi-prosumer in the industrial park based on distributed approximate dynamic programming (ADP), the slope calculation is combined with the PR-ADMM algorithm, which only updates the piecewise linear function slope of the electricity trading volume, carbon quota trading volume, and information within the industrial park's prosumers and consumers.

[0141] The physical meaning of the sampled estimate is the marginal impact of the current state on the value function. By analyzing the changes in the value function after adding a disturbance, the source of the marginal cost can be identified, and the marginal impact can be calculated in a distributed manner based on local information to obtain the sampled estimate. This is applied to energy storage... E and main carbon emissions We will discuss and analyze the sampled estimates of the slope of the piecewise linear function.

[0142] The disturbance is a small disturbance, with an amplitude much smaller than the normal fluctuation range of the state, and will not cause a step change in the operating state of the equipment. For example, it is a small change of ±1% to ±5% of the current state value. By applying this type of disturbance to the energy storage capacity and carbon emission status respectively, the slope sampling estimate of the corresponding piecewise linear function can be calculated in a distributed manner based on the finite difference approximation.

[0143] (1) When energy storage from Become During the process A small perturbation in the state There will be two possible scenarios regarding changes in the operation of the multi-product consumption model in the park: Figure 7 As shown.

[0144] 1) When When the charging and discharging power of the energy storage is not affected, it will not affect the decision-making of the main body. t The operating costs during the period will not change due to disturbances, and the electricity trading volume, carbon allowance trading volume, and status of the other entities will remain unchanged. right The marginal impact is equivalent to the impact on The marginal impact can be expressed using a differential equation as:

[0145] In the formula, For the first k In this iteration, energy storage devices b During the period t Sampling estimate of the slope of a linear function; For the first k In the -1 iteration, energy storage devices b During the period t and segmentation d The slope of the slope; For the first k In this iteration, energy storage devices b During the period t and segmentation d The amount of energy stored below; For the first kIn this iteration, energy storage devices b During the period t The amount of energy stored.

[0146] 2) When When the charging and discharging power of energy storage is affected, the disturbance to the energy storage is equivalent to the disturbance to its charging and discharging power. Based on source-load power balance constraints and carbon emission flow models, this disturbance will impact the decision-making of multiple stakeholders within the region. This situation can be addressed by calculating the local marginal price (LMP) of the node where energy storage b is located, based on the Bellman optimality principle, to obtain a sampled estimate of the slope of the piecewise linear function. Given that energy storage has the characteristic of extending carbon emission conservation from a single time period to carbon emission conservation throughout a trading cycle, then energy storage... b Changes in charging and discharging power will lead to changes in carbon emissions during the current period, energy storage b The LMP of the node will contain the carbon valence component, which is equal to the Lagrange multiplier corresponding to the electricity trading volume obtained by the PR-ADMM algorithm in the electricity-carbon coupled system. Lagrange multipliers corresponding to carbon quota trading volume Coupling:

[0147]

[0148] In the formula, For the first k Energy storage during the next iteration b During the period t Output power; For energy storage b The marginal electricity price at the node where it is located; For energy storage b The node in the time period t The carbon potential.

[0149] (2) When the main carbon emissions from Become During the process A small perturbation in the state The changes in the operation of multi-producer-consumer entities can be divided into two categories based on whether they affect carbon emissions, such as... Figure 8 As shown.

[0150] 1) When If it does not affect the generation of carbon emissions within the park, it will not affect the decision-making of the main entity. t The operating costs during the period will not change due to disturbances, and the electricity trading volume, carbon allowance trading volume, and status of the other entities will remain unchanged. right The marginal impact is equivalent to the impact on The marginal impact can be expressed using a differential equation as:

[0151] In the formula, For the first k In the next iteration, the main body i During the period t Sampling estimate of the slope of a linear function; For the first k In the -1st iteration, the main body i During the period t and segmentation d The slope of the slope; as the main body i During the period t and segmentation d Carbon emissions; For the first k In the next iteration, the main body i During the period t Carbon emissions.

[0152] 2) When When the process of carbon emission generation within the industrial park is affected, the disturbance to carbon emissions is equivalent to the carbon-producing units. g The power output disturbance, based on source-load power balance constraints and carbon emission flow models, will impact the decision-making of multiple stakeholders within the region. This situation can be addressed by calculating carbon-generating units based on Bellman optimality principles. g The LMP at the node obtains a sampled estimate of the slope of the piecewise linear function, which is equal to the Lagrange multiplier corresponding to the electricity trading volume obtained by the PR-ADMM algorithm convergence in the electricity-carbon coupling system. Lagrange multipliers corresponding to carbon quota trading volume Coupling:

[0153]

[0154] In the formula, For the first k Carbon-generating units during the next iteration g During the period t Output power; For carbon-producing units g The marginal electricity price at the node where it is located; For carbon-producing units g The node in the time period t The carbon potential.

[0155] Furthermore, the slopes of each energy storage capacity and the main carbon emission are independent of each other, thus allowing for parallel updates and training, which can significantly improve the efficiency of value function training.

[0156] Step 6.3: Solve the real-time optimization decision-making model of the multi-prosumer cooperative game in the park by using the method of day-ahead offline training and intraday real-time decision-making; like Figure 2 As shown, the specific steps include: (1) Set initial values ​​and input data such as load curves, new energy day-ahead forecast curves, line parameters, unit parameters, and initial carbon quotas within the park; (2) Initialize the slope of each segment in the minimum total running cost function expression after modeling based on the equivalent value of the convex piecewise linear function approximation function, and initialize the number of training iterations. k =1; (3) Set the total number of training iterations K The training scene set is generated using the Latin hypercube sampling method. (4) Initialize the first t The duration of each training session. t =1; (5) Determine the first k During this training session, the system was in time period t The running status; (6) Solve the subproblem of maximizing cooperation benefits, initialize the parameters in the PR-ADMM algorithm, and update the trading volume of electricity and carbon quotas among the subjects according to the objective function and constraint conditions; (7) Determine the convergence condition of the real-time cooperation benefit maximization subproblem. If all subjects can simultaneously satisfy the convergence condition, terminate the iteration process, output the optimal cost and optimal transaction volume in the decision result, and use this as the initial value of the energy payment subproblem. Otherwise, update the multiplier and return to step (2). (8) Solve the real-time energy payment subproblem and calculate the transaction price corresponding to the transaction volume of each producer and consumer in the park in turn; (9) Determine the convergence condition of the real-time energy payment subproblem. If the convergence condition can be satisfied at the same time, terminate the iteration process, output the electricity and carbon quota trading prices in the decision results, otherwise update the multiplier and return to step (3). (10) Update the slope of each segment in the minimum total operating cost function expression, i.e., the slope of the ADP algorithm value; (11) Update time, t = t +1, if the time boundary has not been exceeded. T Then return to step (5); otherwise, proceed to the next step. (12) Update the number of training iterations.k = k +1, if the number of training iterations K is not exceeded, return to step (4); otherwise, proceed to the next step. (13) Output the slope of the trained value function and enter the real-time optimization decision-making process; Distributed intraday real-time optimization decision-making mainly employs a real-time optimization decision-making method for multiple prosumers in the industrial park based on the PR-ADMM algorithm to make decisions based on the actual intraday state. The specific steps are as follows: (14) Input the slope of the value function after training; (15) Initial decision time, t =1; (16) Determine the system during the time period t The operational status of the park is determined by using a distributed real-time optimization decision-making framework based on PR-ADMM to obtain the decision results of the park's multi-prosumer-consumer. (17) Determine if the event-driven criteria are met. If so, execute the optimization decision, wait to enter the next time period, and update the time. t = t +1, otherwise, update the time directly without executing the optimization decision; (18) Determine whether the time boundary has been exceeded. If it has not been exceeded, return to step (16). Otherwise, the process ends and the real-time decision results of the multi-product consumer in the park are output.

[0157] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0158] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0159] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0160] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0161] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A low-carbon response trading method for multi-producer-consumer transactions in industrial parks considering the electricity-carbon market, characterized in that, The method includes the following steps: Step 1: Construct a carbon emission flow model for the power grid and design an electricity-to-carbon ratio model for energy storage devices; Step 2: Based on the carbon emission flow model and the electricity-carbon ratio model, construct a real-time decision-making model for multi-producer-consumer electricity-carbon coordination in the industrial park with the goal of minimizing total operating costs; Step 3: Reconstruct the real-time decision-making model for multi-producer-consumer electricity-carbon collaboration in the park based on Markov decision process, and introduce event-driven criteria; Step 4: Decompose the global optimization objective function in the reconstructed real-time decision-making model of multi-producer-consumer electricity-carbon collaboration in the park into the sum of the operating cost of the current period and the optimal decision cost of the subsequent stage, and obtain the reconstructed objective function applicable to sequential decision-making. Step 5: Based on the reconstructed objective function, construct a real-time optimized transaction decision model for multi-prosumer cooperative game in the park; Step 6: Solve the real-time optimization decision model of the multi-prosumer cooperative game in the park to obtain the real-time decision results of each prosumer; determine whether each prosumer meets the event-driven criterion. If it does, execute the real-time decision result; otherwise, maintain the decision result of the previous moment and wait for the next round of optimization to be triggered.

2. The low-carbon response trading method for multi-producer-consumer in industrial parks considering the electricity-carbon market as described in claim 1, characterized in that: In step 1, the energy storage device's electricity-to-carbon ratio model is specifically as follows: in, , These are energy storage devices e During the period t , t -1 is the carbon-to-electricity ratio of energy storage devices; , These are energy storage devices e During the period t -1、 t Available capacity; For energy storage devices e During the period t The amount of carbon emissions added; For energy storage devices e During the period t The amount of carbon emissions released; For energy storage devices e During the period t The charging power; For energy storage devices e The node in the time period t The node carbon emission intensity; For time intervals; For energy storage devices e During the period t The discharge power; For energy storage devices e When used as a power generation device during a certain period of time t Source-side carbon emission intensity; For energy storage devices e The discharge efficiency.

3. The low-carbon response trading method for multi-producer-consumer in industrial parks considering the electricity-carbon market as described in claim 1, characterized in that: In step 2, the total operating cost includes electricity purchase cost, wind and solar curtailment cost, load shedding cost, electricity market transaction cost, and carbon market transaction cost; the multi-producer-consumer electricity-carbon collaborative real-time decision-making model in the park has constraints including power balance constraints, wind and solar power output constraints, external grid power purchase and sale constraints, energy storage constraints, electricity trading volume constraints, carbon quota trading volume constraints, and carbon emission constraints. The carbon emission constraints are specifically as follows: in, and The main body i During the period t and time period Carbon emissions; Carbon potential of purchased electricity as the main body i and main body j During the period t carbon quota trading volume I For the set of all subjects, as the main body i During the period t Electricity purchased from the power grid.

4. The low-carbon response trading method for multi-producer-consumer in industrial parks considering the electricity-carbon market as described in claim 1, characterized in that: In step 3, the real-time collaborative decision-making model for electricity and carbon emissions among multiple producers and consumers in the park, reconstructed based on the Markov decision process, specifically includes: using the energy storage capacity, carbon emissions, actual renewable energy output, and load power of each entity in the park as state variables, and the energy storage charging and discharging power, electricity trading volume, carbon quota trading volume, and actual renewable energy consumption as decision variables, combined with external information, to construct a state transition equation to characterize the dynamic operating characteristics of the system; the external information includes the prediction error and day-ahead forecast value of renewable energy, and the prediction error and day-ahead forecast value of load.

5. The low-carbon response trading method for multi-producer-consumer in industrial parks considering the electricity-carbon market as described in claim 3, characterized in that: In step 3, the event-driven criterion is specifically as follows: in, as the main body i During the period t In terms of carbon emissions; as the main body i The carbon emission threshold.

6. The low-carbon response trading method for multi-producer-consumer in industrial parks considering the electricity-carbon market as described in claim 1, characterized in that: In step 4, the reconstructed objective function is specifically as follows: in, From t The minimum total running cost function from the start of the trading session to the end of the transaction; , They are respectively t Decision variables and state variables for different time periods; for t Time period in state ,decision making Current operating costs for the current period; In order to be in t Make decisions during the time period The new state that the system then enters; Indicates a new state The minimum cumulative operating cost from start to finish of the transaction is the optimal decision cost for subsequent stages.

7. The low-carbon response trading method for multi-producer-consumer in industrial parks considering the electricity-carbon market as described in claim 6, characterized in that: The real-time optimization transaction decision model for multi-producer-consumer cooperative game in the park includes a cooperative benefit maximization sub-model and an energy transaction payment sub-model. The cooperative benefit maximization sub-model is used to determine the optimal electricity and carbon quota trading volume among the various entities, and the energy transaction payment sub-model is used to determine the optimal electricity and carbon quota trading price among the various entities. The two sub-models are solved in a distributed iterative manner using an augmented Lagrangian function.

8. The low-carbon response trading method for multi-producer-consumer in industrial parks considering the electricity-carbon market as described in claim 7, characterized in that: Step 6 involves solving the real-time optimization decision-making model for multi-prosumer cooperative game in the park, which includes two stages: day-ahead offline training and day-ahead real-time decision-making. Specifically: During the recent offline training phase: The PR-ADMM algorithm is used to solve the cooperation benefit maximization sub-model and the real-time energy transaction payment sub-model to obtain the optimal electricity and carbon quota trading volume and the optimal electricity and carbon quota trading price among the participants. In the reconstructed objective function, the high-dimensional state variables related to energy flow are aggregated into energy storage capacity, and the high-dimensional state variables related to carbon flow are aggregated into carbon emissions. Based on the approximation function of the convex piecewise linear function, the equivalent value of the optimal decision cost in the subsequent stage of the reconstructed objective function is modeled to obtain the reconstructed objective function after the equivalent value modeling. Based on the optimal electricity and carbon quota trading volume and the optimal electricity and carbon quota trading price among the aforementioned entities, calculate the slope of each segment in the reconstructed objective function after equivalent value modeling; During the intraday real-time decision-making phase: Based on the slopes of each segment calculated during the offline training phase, the reconstruction objective function after the equivalent value model is constructed and solved to obtain the real-time decision results of each producer and consumer.

9. The low-carbon response trading method for multi-producer-consumer in industrial parks considering the electricity-carbon market as described in claim 8, characterized in that: The reconstructed objective function after modeling the equivalent values ​​is specifically as follows: in, D The total number of segments in a piecewise linear function; and Energy storage b and main body i During the period t No. d The slope of the segment; For energy storage b During the period t No. d Segmented energy storage capacity; as the main body i During the period t No. d Segmented carbon emissions; D The total number of segments in a piecewise linear function; B A collection of energy storage devices.

10. The low-carbon response trading method for multi-producer-consumer in industrial parks considering the electricity-carbon market as described in claim 9, characterized in that: The slope of each segment in the reconstructed objective function after the equivalent value modeling is determined as follows: in, k This represents the number of iterations. , The first k In this iteration, energy storage devices b ,main body i During the period and segmentation The slope of the slope; , The first k In the -1 iteration, energy storage devices b ,main body i During the period and segmentation Lower segment slope; , They were respectively in the second k Energy storage during the next iteration b ,main body i During the period t The sampled estimate of the slope of the lower piecewise linear function; Divide the linear function into segments; To update the step size, and satisfy the following conditions: .

11. The low-carbon response trading method for multi-producer-consumer in industrial parks considering the electricity-carbon market as described in claim 10, characterized in that: The sampled estimate of the slope of the piecewise linear function is calculated as follows: A perturbation is applied to the energy storage state, and the slope sampling estimate of each energy storage device in the piecewise linear function is calculated in two scenarios based on whether the perturbation affects the energy storage charging and discharging power. When the charging and discharging power is not affected, the slope sampling estimate of each energy storage device is calculated as follows: in, For the first k In this iteration, energy storage devices b During the period t and segmentation The sampled estimate of the slope of the linear function under the given conditions; For the first k In the -1 iteration, energy storage devices b During the period t and segmentation d The slope of the slope; For the first k In this iteration, energy storage devices b During the period t and segmentation d The amount of energy stored below; For the first k In this iteration, energy storage devices b During the period t The energy storage capacity; When considering factors affecting charging and discharging power, the slope sampling estimate for each energy storage device is calculated as follows: in, For energy storage b The marginal electricity price at the node where it is located; A perturbation is applied to the carbon emission status, and the slope sampling estimate of each subject in the piecewise linear function is calculated in two scenarios, depending on whether the perturbation affects the generation of carbon emissions within the park. When the generation of carbon emissions within the park is not affected, the slope sampling estimate for each entity is calculated as follows: in, For the first k In the next iteration, the main body i During the period t Sampling estimate of the slope of a linear function; For the first k In the -1st iteration, the main body i During the period t and segmentation d The slope of the slope; as the main body i During the period t and segmentation d Carbon emissions; For the first k In the next iteration, the main body i During the period t Carbon emissions; When influencing carbon emissions within the park, the slope sampling estimate for each entity is calculated as follows: in, For carbon-producing units g The marginal electricity price of the node.

12. A low-carbon response trading system for multi-prosumer entities in industrial parks, considering the electricity-carbon market, based on the method of any one of claims 1-11, comprising a power grid carbon emission flow model construction module, a real-time decision-making model construction module for multi-prosumer entities in industrial parks (electricity-carbon collaborative model), a real-time decision-making model reconstruction module for multi-prosumer entities in industrial parks (electricity-carbon collaborative model), an objective function reconstruction module, a real-time optimization trading decision-making model construction module for multi-prosumer entities in industrial parks (cooperative game), and a solution module for the real-time optimization trading decision-making model for multi-prosumer entities in industrial parks (cooperative game), characterized in that: A module for building a carbon emission flow model for power grids is used to construct a carbon emission flow model for power grids and design an energy storage device's carbon-to-electricity ratio model. The module for constructing a real-time decision-making model for multi-producer-consumer electricity-carbon coordination in industrial parks, based on the carbon emission flow model and the electricity-carbon ratio model, constructs a real-time decision-making model for multi-producer-consumer electricity-carbon coordination in industrial parks with the goal of minimizing total operating costs. The module for reconstructing the real-time decision-making model of multi-producer-consumer electricity-carbon collaboration in the industrial park reconstructs the model based on the Markov decision process and introduces event-driven criteria. The objective function reconstruction module decomposes the global optimization objective function in the real-time decision-making model of multi-producer-consumer electricity-carbon collaboration in the park into a recursive combination of the current operating cost and the optimal decision cost in the subsequent stage, and obtains a reconstructed objective function suitable for sequential decision-making. The module for constructing a real-time optimized transaction decision model for multi-prosumer cooperative game in the industrial park is based on the reconstructed objective function to construct a real-time optimized transaction decision model for multi-prosumer cooperative game in the industrial park. The module for solving the real-time optimization transaction decision model of the multi-prosumer cooperative game in the park solves the real-time optimization decision model of the multi-prosumer cooperative game in the park and obtains the real-time decision results of each prosumer. Determine whether each producer and consumer meets the event-driven criteria. If it does, execute the real-time decision result; otherwise, maintain the decision result from the previous moment and wait for the next round of optimization to be triggered.

13. A terminal, comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the method according to any one of claims 1-11.

14. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method according to any one of claims 1-11.

Images