A multi-garden integrated energy system electric-carbon collaborative optimization method and system

Abstract

This invention discloses a method and system for coordinated optimization of carbon emissions in a multi-park integrated energy system, belonging to the field of integrated energy optimization and scheduling technology. The method includes: constructing a dynamic carbon trading and green certificate flexible offsetting mechanism; building a two-layer game optimization model with the system operator as the leader and multiple park agents as followers; the upper-layer system operator optimization model uses the DQN algorithm to dynamically optimize and publish energy trading prices; the lower-layer park agent optimization model uses the PPO algorithm to perform distributed optimization scheduling of equipment within the park based on the received energy trading prices; improving the Shapley value benefit allocation method based on carbon emission reduction contributions to fairly distribute the total alliance benefits generated by the coordinated operation of multiple parks; and conducting collaborative iterative training of the two-layer game optimization model until game equilibrium is reached, outputting the optimal energy trading price and scheduling strategies for each park. This invention achieves a synergistic improvement in both the overall economic benefits and carbon emission reduction levels of the system.

Description

Technical Field

[0001] This invention belongs to the field of integrated energy optimization and scheduling technology, specifically relating to a method and system for coordinated optimization of electricity and carbon emissions in a multi-park integrated energy system. Background Technology

[0002] The statements herein provide only background information in relation to this invention and do not necessarily constitute prior art.

[0003] Driven by the "dual-carbon" strategic goals, regional integrated energy systems characterized by multi-park interconnection have become a key vehicle for improving energy efficiency, promoting renewable energy consumption, and achieving large-scale emission reduction. Such systems, through the synergy and mutual support of multiple park integrated energy systems, can optimize resource allocation on a larger scale. However, their market-oriented operation faces multiple coupling challenges: First, at the mechanism level, existing research often treats carbon trading costs as fixed or tiered prices and incorporates them into optimization models. For example, it uses a fixed carbon price or tiered carbon prices set according to emission ranges. Such static mechanisms are difficult to sensitively and in real-time reflect the dynamic changes in the supply and demand relationship of carbon emissions in the operation of the energy system. They have limited guiding role in the low-carbon dispatching behavior of market participants and cannot fully realize the potential of the carbon market as a price signal in real-time adjustment of supply and demand and incentivizing deep emission reduction.

[0004] Secondly, at the collaborative model level, each park in a multi-park system acts as an independent stakeholder, engaging in a game between economic interests and emission reduction responsibilities. Traditional centralized optimization scheduling requires collecting all operational data from each park, posing a risk of commercial privacy breaches and placing enormous computational pressure on the central system. While existing game theory-based methods can characterize the interactions between stakeholders, they often face challenges in solving high-dimensional, nonlinear, and multi-time-period coupled decision-making problems, such as slow convergence speed and difficulty in obtaining stable equilibrium solutions, failing to meet the timeliness requirements of online optimization in practical systems.

[0005] Finally, at the incentive compatibility level, there is a lack of an effective allocation mechanism that fairly and transparently links the carbon emission reduction contributions of individual parks to the overall benefits of collaborative operation. This means that parks with large emission reduction contributions may not receive corresponding rewards, weakening their intrinsic motivation to participate in system-level low-carbon collaborative optimization and affecting the long-term stability of the alliance. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the existing technology and provide a method and system for synergistic optimization of carbon emissions in a multi-park integrated energy system. By constructing a two-layer optimization framework that deeply integrates market mechanisms and artificial intelligence algorithms, it effectively solves problems such as conflicts of interest among multiple stakeholders, insufficient carbon market incentives, and data privacy and security, thereby achieving a synergistic improvement in the overall economic benefits and carbon emission reduction level of the system.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solution: In a first aspect, the technical solution of the present invention provides a method for coordinated optimization of electricity and carbon emissions in a multi-park integrated energy system, comprising: A dynamic carbon trading and green certificate flexible offset mechanism is established; the carbon trading benchmark price is dynamically adjusted according to the market supply and demand relationship of carbon emission quotas, and the offset coefficient of green certificates against carbon emissions is configured as a dynamic function related to the proportion of renewable energy output in the park. A two-layer game optimization model is constructed with the system operator as the leader and multiple park agents as followers. The upper-layer system operator optimization model adopts the DQN algorithm to maximize its own operational economic benefits and system carbon emission reduction benefits, and dynamically optimizes and publishes energy trading prices. The lower-layer park agent optimization model adopts the PPO algorithm to perform distributed optimization scheduling of equipment within the park based on the received energy trading prices, with the goal of minimizing its own total operating costs. Based on the improved Shapley value benefit distribution method based on carbon emission reduction contribution, the total alliance benefits generated by the collaborative operation of multiple parks are distributed fairly. The two-layer game optimization model is trained through collaborative iteration. Through the feedback loop of price signals and scheduling plans, the game equilibrium is reached, and the optimal energy trading price and scheduling strategy for each park are output.

[0008] In at least one embodiment, the carbon trading benchmark price is expressed as:

[0009] In the formula, express The benchmark price for carbon trading during a specific period; Indicates the initial time-based benchmark price for carbon trading; express Carbon emission demand during a given period; express Carbon emission supply during a given period; This represents the adjustment coefficient.

[0010] In at least one embodiment, the carbon emission offset factor of green certificates is expressed as:

[0011] In the formula, express The offset coefficient of carbon emissions for time-limited green certificates; This indicates the offsetting factor of green certificates against carbon emissions at the initial time. express The percentage of renewable energy output in the industrial park during different time periods; express The percentage of electricity load in the park during different time periods; This represents the sensitivity coefficient.

[0012] In at least one embodiment, in the upper-level system operator optimization model, the state space includes historical and current energy trading prices, total system load demand, and total predicted renewable energy output; the action space is the adjustment range of green electricity trading prices and conventional electricity trading prices; and the reward function is the weighted sum of the system operator's economic benefits and the system's carbon emission reduction revenue.

[0013] In at least one embodiment, the constraints of the upper-layer system operator optimization model include electricity price constraints and supply-demand balance constraints.

[0014] In at least one embodiment, the lower-level park agent optimization model includes the energy trading price published by the system operator, the park's local electricity / heat / cooling load data, the real-time output of renewable energy and the state of charge of energy storage devices; the action space includes the output combination and start-up / shutdown strategies of gas turbines, photovoltaic systems, energy storage devices and adjustable loads within the park; and its reward function is the negative value of the park's total operating cost; wherein, the park's total operating cost includes energy purchase cost, equipment operation and maintenance cost, carbon trading cost and green certificate trading revenue.

[0015] In at least one embodiment, the improved Shapley value benefit allocation method based on carbon emission reduction contribution is specifically expressed as follows:

[0016] In the formula, Indicates the park The final distribution of profits; This represents the allocation amount based on the traditional Shapley value; This indicates the carbon emission reduction contribution weighting factor, which is related to the industrial park. The carbon emission reduction is positively correlated with the penetration rate of renewable energy.

[0017] Secondly, the technical solution of the present invention also provides a multi-park integrated energy system for coordinated optimization of electricity and carbon emissions, comprising: The dynamic adjustment module is configured to: construct a dynamic carbon trading and green certificate flexible offset mechanism; wherein, the carbon trading benchmark price is dynamically adjusted according to the market supply and demand relationship of carbon emission quotas, and the offset coefficient of green certificates against carbon emissions is configured as a dynamic function related to the proportion of renewable energy output in the park. The model building module is configured to construct a two-layer game optimization model with the system operator as the leader and multiple park agents as followers. The upper-layer system operator optimization model adopts the DQN algorithm to maximize its own operational economic benefits and system carbon emission reduction benefits, and dynamically optimizes and publishes energy trading prices. The lower-layer park agent optimization model adopts the PPO algorithm to perform distributed optimization scheduling of equipment within the park based on the received energy trading prices, with the goal of minimizing its own total operating costs. The allocation module is configured to: improve the Shapley value benefit allocation method based on carbon emission reduction contribution, and fairly allocate the total alliance benefits generated by the collaborative operation of multiple parks; The iterative training module is configured to: conduct collaborative iterative training of the two-layer game optimization model, and through the feedback loop of price signals and scheduling plans, until the game equilibrium is reached, and output the optimal energy trading price and scheduling strategy for each park.

[0018] Thirdly, the technical solution of the present invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps in the multi-park integrated energy system electric carbon synergistic optimization method described in the first aspect.

[0019] Fourthly, the present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the multi-park integrated energy system electric carbon synergistic optimization method described in the first aspect.

[0020] The beneficial effects of the above-described technical solution of the present invention are as follows: The multi-park integrated energy system electricity-carbon collaborative optimization method of this invention effectively guides the energy consumption structure towards low-carbon transformation through dynamic carbon pricing and flexible green certificate mechanisms. Employing a deep reinforcement learning framework, each park can make optimization decisions using only local information, eliminating the need to share sensitive data and solving information privacy issues. The hybrid reinforcement learning framework of DQN and PPO effectively handles high-dimensional state and action spaces, significantly improving convergence speed. By introducing the Shapley value method for carbon contribution weighting, parks with greater emission reduction contributions receive more benefits, effectively improving allocation fairness and enhancing the stability of the alliance. Attached Figure Description

[0021] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0022] Figure 1 This is a schematic diagram of a multi-park integrated energy system electricity-carbon synergistic optimization method disclosed in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the dynamic carbon trading and green certificate flexible offset mechanism disclosed in Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of the two-layer game optimization model disclosed in Embodiment 1 of the present invention; Figure 4 This is a schematic diagram of the electricity-carbon synergy of the integrated energy system in the park disclosed in Embodiment 1 of the present invention; Figure 5 This is a comparison diagram of the iterative optimization process of the method proposed in this invention and the traditional algorithm disclosed in Embodiment 1 of this invention; Figure 6 This is a schematic diagram showing the comparison between the total system cost and carbon emissions disclosed in Embodiment 1 of the present invention; Figure 7 This is a schematic diagram of the distribution results of park benefits disclosed in Embodiment 1 of the present invention. Detailed Implementation

[0023] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0024] As described in the background section, the purpose of this invention is to overcome the shortcomings of the existing technologies and provide a method and system for synergistic optimization of carbon emissions in multi-park integrated energy systems. By constructing a two-layer optimization framework that deeply integrates market mechanisms and artificial intelligence algorithms, this invention effectively solves problems such as conflicts of interest among multiple stakeholders, insufficient carbon market incentives, and data privacy and security, thereby achieving a synergistic improvement in the overall economic benefits and carbon emission reduction levels of the system.

[0025] Example 1 In a typical embodiment of the present invention, such as Figures 1 to 7 As shown in the figure, this embodiment discloses a method for coordinated optimization of electricity and carbon emissions in a multi-park integrated energy system, which specifically includes the following steps: S1. Construct a dynamic carbon trading and green certificate flexible offset mechanism; wherein, the carbon trading benchmark price is dynamically adjusted according to the market supply and demand relationship of carbon emission quotas, and the offset coefficient of green certificates for carbon emissions is configured as a dynamic function related to the proportion of renewable energy output in the park; S2. A two-layer game optimization model is constructed with the system operator as the leader and multiple park agents as followers. The upper-layer system operator optimization model adopts the DQN algorithm to maximize its own operational economic benefits and system carbon emission reduction benefits, and dynamically optimizes and publishes energy trading prices. The lower-layer park agent optimization model adopts the PPO algorithm to perform distributed optimization scheduling of equipment within the park based on the received energy trading prices, with the goal of minimizing its own total operating costs. S3. Improve the Shapley value benefit distribution method based on carbon emission reduction contribution to fairly distribute the total alliance benefits generated by the collaborative operation of multiple parks; S4. The two-layer game optimization model is trained collaboratively and iteratively. Through the feedback loop of price signals and scheduling plans, the game equilibrium is reached, and the optimal energy trading price and scheduling strategy for each park are output.

[0026] The following section provides a detailed explanation of the above-mentioned method for co-optimizing the electricity and carbon emissions of a multi-park integrated energy system, using specific implementation methods as examples.

[0027] S1. Construct a dynamic carbon trading and green certificate flexible offset mechanism; wherein, the carbon trading benchmark price is dynamically adjusted according to the market supply and demand relationship of carbon emission allowances, and the offset coefficient of green certificates against carbon emissions is configured as a dynamic function related to the proportion of renewable energy output in the park, such as... Figure 2 As shown.

[0028] S11. Construct a dynamic carbon trading model.

[0029] In this step, the carbon trading benchmark price is dynamically adjusted based on the market supply and demand relationship of carbon emission allowances, specifically expressed as follows:

[0030] In the formula, express The benchmark price for carbon trading during a specific period; Indicates the initial time-based benchmark price for carbon trading; express Carbon emission demand during a given period; express Carbon emission supply during a given period; This represents the adjustment coefficient, and its value range is... .when At that time, the carbon trading benchmark price will rise, incentivizing the park to reduce the procurement of high-carbon energy.

[0031] S12. Construct a green certificate-carbon emission elasticity offset model.

[0032] In this step, the carbon emission offset coefficient of green certificates is designed as a dynamic function related to the proportion of renewable energy output in the industrial park. This dynamically links the carbon emission offset coefficient of green certificates with the proportion of renewable energy output in the park, allowing parks with a high proportion of green electricity consumption to obtain higher carbon offset benefits. Specifically, this can be expressed as:

[0033] In the formula, express The offset coefficient of carbon emissions for time-limited green certificates; This indicates the offsetting factor of green certificates against carbon emissions at the initial time. express The percentage of renewable energy output in the industrial park during different time periods; express The percentage of electricity load in the park during different time periods; This represents the sensitivity coefficient.

[0034] S2. A two-layer game optimization model is constructed with the system operator as the leader and multiple park agents as followers. The upper-layer system operator optimization model adopts the DQN algorithm, aiming to maximize its own operational economic benefits and system carbon emission reduction revenue, and dynamically optimizes and publishes energy trading prices. The lower-layer park agent optimization model adopts the PPO algorithm, based on the received energy trading prices, aiming to minimize its own total operating costs, and performs distributed optimization scheduling of equipment within the park, such as... Figure 3 As shown.

[0035] S21. Construct an upper-layer system operator optimization model.

[0036] In this step, the system operator, as the leader, constructs an upper-level system operator optimization model with the goal of maximizing its own operational (ISO) economic benefits and system carbon emission reduction benefits. This model dynamically optimizes and publishes energy trading prices, specifically as follows:

[0037] In the formula, This represents the total benefits received by ISO. Indicates the operating revenue of the power system; Indicates the operating cost of the power system; Indicates the reduction in carbon emissions; This represents the carbon efficiency conversion factor, expressed in yuan / ton.

[0038] In this step, the upper-layer system operator optimization model is a sequential decision-making model based on the Deep Q-Network (DQN) algorithm, and its decision variables include the green electricity trading price. and conventional electricity transaction prices .

[0039] In this step, the core of the upper-level system operator optimization model is to enable the system operator, as a market leader, to maximize its overall revenue through dynamic pricing strategies, and to drive the low-carbon operation of the lower-level park agent optimization model. Its state space includes historical and current energy trading prices, total system load demand, and total projected renewable energy output; its action space is the green electricity trading price. Compared with conventional electricity transaction prices The adjustment range; the reward function is the weighted sum of the system operator's economic benefits and the system's carbon emission reduction revenue. .

[0040] In this step, the constraints of the upper-level system operator optimization model include electricity price constraints and supply-demand balance constraints. Specifically, the electricity price constraint is expressed as follows:

[0041] In the formula, express The price of green electricity during certain time periods; express Green electricity purchase price during certain periods; This indicates the highest transaction price in the green electricity market.

[0042] The supply and demand balance constraint is specifically expressed as:

[0043] In the formula, This indicates the total generating capacity of the green electricity buyer; This indicates the total generating capacity of the green electricity seller.

[0044] S22. Construct a lower-level park agent optimization model.

[0045] In this step, each park acts as a follower, constructing a lower-level park agent optimization model based on the energy trading prices published by the upper-level system operator's optimization model, with the goal of minimizing its own total operating costs. This model enables distributed optimization scheduling of equipment within the park, specifically represented as follows:

[0046] In the formula, Indicates the total cost; Indicates energy transaction costs; Indicates system operating costs; Indicates the cost of natural gas consumption; Indicates carbon emissions; express Benchmark price for carbon trading during specific time periods.

[0047] In this step, the optimization model for each lower-level park agent is a distributed decision-making model based on the Proximal Policy Optimization (PPO) algorithm, and its decision variables include equipment output variables. and energy storage charging and discharging strategies .

[0048] In this step, the core objective of the lower-level park agent optimization model is to optimize the output plan of each device within the park based on the energy trading prices published by the upper-level system operator optimization model and local state information, in order to minimize the park's total operating cost. Its state space includes the energy trading prices published by the system operator (green electricity purchase and sale prices and conventional electricity purchase and sale prices), the park's local electricity / heat / cooling load data, real-time renewable energy output, and the state of charge of energy storage devices; its action space includes the output combinations and start-up / shutdown strategies of gas turbines, photovoltaic systems, energy storage devices, and adjustable loads within the park; and its reward function is the park's total operating cost. The negative value indicates that the total operating cost includes energy purchase cost, equipment operation and maintenance cost, carbon trading cost, and green certificate trading revenue.

[0049] S3. Improve the Shapley value benefit distribution method based on carbon emission reduction contribution to fairly distribute the total alliance benefits generated by the collaborative operation of multiple parks.

[0050] In the cooperative alliance model, an improved Shapley value method is used to fairly distribute the total alliance benefits generated by the collaborative operation of multiple parks, effectively incentivizing each park to actively pursue low-carbon transformation. The improved Shapley value method modifies the traditional Shapley value by introducing a carbon emission reduction contribution weighting factor, allowing parks with greater emission reduction contributions to receive more benefits, thereby improving the fairness of distribution and the stability of the alliance. The calculation formula is as follows:

[0051] In the formula, Indicates the park The final distribution of profits; This represents the allocation amount based on the traditional Shapley value; This indicates the carbon emission reduction contribution weighting factor, which is related to the industrial park. Carbon emission reduction and renewable energy penetration are positively correlated, with values ​​ranging from [value range missing]. .

[0052] S4. The two-layer game optimization model is trained collaboratively and iteratively. Through the feedback loop of price signals and scheduling plans, the game equilibrium is reached, and the optimal energy trading price and scheduling strategy for each park are output.

[0053] In this step, the upper-level system operator optimization model using the DQN algorithm and the lower-level park agent optimization model using the PPO algorithm undergo collaborative iterative training. The upper-level system operator optimization model publishes energy trading price strategies, and each lower-level park agent optimization model performs local optimization based on the received energy trading prices and feeds back its electricity dispatch plan. The upper-level system operator optimization model then updates its pricing strategy based on the feedback. This process is repeated until the rate of change of the total system cost is less than a set threshold (e.g., ...). This is considered to be the achievement of game equilibrium. The algorithm converges and outputs the optimal energy trading price and scheduling strategy for each park.

[0054] The following example, using a multi-park integrated energy system comprising three interconnected industrial parks in a certain region, verifies the proposed method for coordinated optimization of electricity and carbon emissions in multi-park integrated energy systems. The equipment within the parks includes photovoltaic systems, gas turbines, energy storage devices, and various loads, such as… Figure 4 As shown.

[0055] Step 1: Parameter settings.

[0056] Dynamic carbon trading parameters: benchmark carbon price Yuan / ton, adjustment coefficient .

[0057] Green Certificate Flexible Offset Parameters: Benchmark Offset Coefficient =0.6, sensitivity coefficient .

[0058] Reinforcement learning parameters: DQN learning rate PPO discount factor .

[0059] Carbon emission reduction contribution weighting: In the Shapley value method, a weighting factor is set for the contribution of carbon emission reduction. .

[0060] Step 2: Construct a two-layer game optimization model.

[0061] In a Python environment, the DQN and PPO algorithm networks are built using the TensorFlow framework.

[0062] Upper-level system operator optimization model: The state space dimension includes electricity prices, total load forecasts, and total renewable energy output forecasts for each time period within 24 hours; the action space includes the adjustment range of green electricity trading prices and conventional electricity trading prices.

[0063] Lower-level park agent optimization model: The state space of each park agent includes the 24-hour electricity price sequence, local load, photovoltaic output and energy storage SOC; the action space is the 24-hour output plan and start-stop strategy of each controllable device.

[0064] Set up comparison scenarios: Scenario A, which adopts the independent operation + static carbon trading method, and Scenario B, which adopts the multi-park integrated energy system electricity-carbon synergistic optimization method disclosed in this embodiment.

[0065] Step 3: Results Analysis After simulation iterations, the following results were obtained: Economic efficiency: The total operating cost of the alliance in scenario B is RMB 49,491.59, which is 14.8% lower than RMB 58,129.30 in scenario A.

[0066] Low carbon footprint: The total carbon emissions for scenario B are 94.85 tons, a reduction of 24.8% compared to 126.11 tons for scenario A.

[0067] Algorithm performance: such as Figure 5 As shown, the proposed DQN-PPO hybrid algorithm converges after 18 iterations, while the traditional Harris Eagle Optimization (HHO) algorithm requires 25 iterations, improving the convergence speed by approximately 35%.

[0068] Profit distribution: such as Figure 6 and Figure 7 As shown, after applying the carbon contribution-weighted Shapley value method, parks with high renewable energy penetration (such as...) The benefits increased by 28%, and the Gini coefficient for the distribution of alliance benefits decreased from 0.32 to 0.24, significantly improving fairness.

[0069] Example 2 In a typical embodiment of the present invention, this embodiment discloses a multi-park integrated energy system for coordinated optimization of electricity and carbon emissions, specifically including: The dynamic adjustment module is configured to: construct a dynamic carbon trading and green certificate flexible offset mechanism; wherein, the carbon trading benchmark price is dynamically adjusted according to the market supply and demand relationship of carbon emission quotas, and the offset coefficient of green certificates against carbon emissions is configured as a dynamic function related to the proportion of renewable energy output in the park. The model building module is configured to construct a two-layer game optimization model with the system operator as the leader and multiple park agents as followers. The upper-layer system operator optimization model adopts the DQN algorithm to maximize its own operational economic benefits and system carbon emission reduction benefits, and dynamically optimizes and publishes energy trading prices. The lower-layer park agent optimization model adopts the PPO algorithm to perform distributed optimization scheduling of equipment within the park based on the received energy trading prices, with the goal of minimizing its own total operating costs. The allocation module is configured to: improve the Shapley value benefit allocation method based on carbon emission reduction contribution, and fairly allocate the total alliance benefits generated by the collaborative operation of multiple parks; The iterative training module is configured to: conduct collaborative iterative training of the two-layer game optimization model, and through the feedback loop of price signals and scheduling plans, until the game equilibrium is reached, and output the optimal energy trading price and scheduling strategy for each park.

[0070] Example 3 In a typical embodiment of the present invention, this embodiment provides a computer-readable storage medium storing a computer program thereon. When executed by a processor, the program implements the steps in the multi-park integrated energy system electricity-carbon synergistic optimization method described in Embodiment 1. These steps include: S1. Construct a dynamic carbon trading and green certificate flexible offset mechanism; wherein, the carbon trading benchmark price is dynamically adjusted according to the market supply and demand relationship of carbon emission quotas, and the offset coefficient of green certificates for carbon emissions is configured as a dynamic function related to the proportion of renewable energy output in the park; S2. A two-layer game optimization model is constructed with the system operator as the leader and multiple park agents as followers. The upper-layer system operator optimization model adopts the DQN algorithm to maximize its own operational economic benefits and system carbon emission reduction benefits, and dynamically optimizes and publishes energy trading prices. The lower-layer park agent optimization model adopts the PPO algorithm to perform distributed optimization scheduling of equipment within the park based on the received energy trading prices, with the goal of minimizing its own total operating costs. S3. Improve the Shapley value benefit distribution method based on carbon emission reduction contribution to fairly distribute the total alliance benefits generated by the collaborative operation of multiple parks; S4. The two-layer game optimization model is trained collaboratively and iteratively. Through the feedback loop of price signals and scheduling plans, the game equilibrium is reached, and the optimal energy trading price and scheduling strategy for each park are output.

[0071] Example 4 In a typical embodiment of the present invention, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the multi-park integrated energy system electricity-carbon synergistic optimization method described in Embodiment 1. These steps include: S1. Construct a dynamic carbon trading and green certificate flexible offset mechanism; wherein, the carbon trading benchmark price is dynamically adjusted according to the market supply and demand relationship of carbon emission quotas, and the offset coefficient of green certificates for carbon emissions is configured as a dynamic function related to the proportion of renewable energy output in the park; S2. A two-layer game optimization model is constructed with the system operator as the leader and multiple park agents as followers. The upper-layer system operator optimization model adopts the DQN algorithm to maximize its own operational economic benefits and system carbon emission reduction benefits, and dynamically optimizes and publishes energy trading prices. The lower-layer park agent optimization model adopts the PPO algorithm to perform distributed optimization scheduling of equipment within the park based on the received energy trading prices, with the goal of minimizing its own total operating costs. S3. Improve the Shapley value benefit distribution method based on carbon emission reduction contribution to fairly distribute the total alliance benefits generated by the collaborative operation of multiple parks; S4. The two-layer game optimization model is trained collaboratively and iteratively. Through the feedback loop of price signals and scheduling plans, the game equilibrium is reached, and the optimal energy trading price and scheduling strategy for each park are output.

[0072] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for coordinated optimization of electricity and carbon emissions in a multi-park integrated energy system, characterized in that, include: A dynamic carbon trading and green certificate flexible offset mechanism is established; the carbon trading benchmark price is dynamically adjusted according to the market supply and demand relationship of carbon emission quotas, and the offset coefficient of green certificates against carbon emissions is configured as a dynamic function related to the proportion of renewable energy output in the park. A two-layer game optimization model is constructed with the system operator as the leader and multiple park agents as followers. The upper-layer system operator optimization model adopts the DQN algorithm to maximize its own operational economic benefits and system carbon emission reduction benefits, and dynamically optimizes and publishes energy trading prices. The lower-layer park agent optimization model adopts the PPO algorithm to perform distributed optimization scheduling of equipment within the park based on the received energy trading prices, with the goal of minimizing its own total operating costs. Based on the improved Shapley value benefit distribution method based on carbon emission reduction contribution, the total alliance benefits generated by the collaborative operation of multiple parks are distributed fairly. The two-layer game optimization model is trained through collaborative iteration. Through the feedback loop of price signals and scheduling plans, the game equilibrium is reached, and the optimal energy trading price and scheduling strategy for each park are output.

2. The method for coordinated optimization of electricity and carbon emissions in a multi-park integrated energy system as described in claim 1, characterized in that, The carbon trading benchmark price is expressed as: In the formula, express The benchmark price for carbon trading during a specific period; Indicates the initial time-based benchmark price for carbon trading; express Carbon emission demand during a given period; express Carbon emission supply during a given period; This represents the adjustment coefficient.

3. The method for coordinated optimization of electricity and carbon emissions in a multi-park integrated energy system as described in claim 1, characterized in that, The carbon emission offset factor of green certificates is expressed as: In the formula, express The offset coefficient of carbon emissions for time-limited green certificates; This indicates the offsetting factor of green certificates against carbon emissions at the initial time. express The percentage of renewable energy output in the industrial park during different time periods; express The percentage of electricity load in the park during different time periods; This represents the sensitivity coefficient.

4. The method for coordinated optimization of electricity and carbon emissions in a multi-park integrated energy system as described in claim 1, characterized in that, In the upper-level system operator optimization model, the state space includes historical and current energy trading prices, total system load demand, and total predicted renewable energy output; the action space is the adjustment range of green electricity trading prices and conventional electricity trading prices; and the reward function is the weighted sum of the system operator's economic benefits and the system's carbon emission reduction revenue.

5. The method for coordinated optimization of electricity and carbon emissions in a multi-park integrated energy system as described in claim 4, characterized in that, The constraints of the upper-level system operator optimization model include electricity price constraints and supply-demand balance constraints.

6. The method for coordinated optimization of electricity and carbon emissions in a multi-park integrated energy system as described in claim 1, characterized in that, In the lower-level park agent optimization model, the state space includes the energy trading prices published by the system operator, the local electricity / heat / cooling load data of the park, the real-time output of renewable energy and the state of charge of energy storage devices; The action space is the output combination and start-up / shutdown strategy of gas turbines, photovoltaic systems, energy storage devices and adjustable loads in the park; its reward function is the negative value of the total operating cost of the park; wherein, the total operating cost of the park includes energy purchase cost, equipment operation and maintenance cost, carbon trading cost and green certificate trading revenue.

7. The method for coordinated optimization of electricity and carbon emissions in a multi-park integrated energy system as described in claim 1, characterized in that, The improved Shapley value benefit allocation method based on carbon emission reduction contribution is specifically expressed as follows: In the formula, Indicates the park The final distribution of profits; This represents the allocation amount based on the traditional Shapley value; This indicates the carbon emission reduction contribution weighting factor, which is related to the industrial park. The carbon emission reduction is positively correlated with the penetration rate of renewable energy.

8. A multi-park integrated energy system for coordinated optimization of electricity and carbon emissions, characterized in that, include: The dynamic adjustment module is configured to: construct a dynamic carbon trading and green certificate flexible offset mechanism; wherein, the carbon trading benchmark price is dynamically adjusted according to the market supply and demand relationship of carbon emission quotas, and the offset coefficient of green certificates against carbon emissions is configured as a dynamic function related to the proportion of renewable energy output in the park. The model building module is configured to construct a two-layer game optimization model with the system operator as the leader and multiple park agents as followers. The upper-layer system operator optimization model adopts the DQN algorithm to maximize its own operational economic benefits and system carbon emission reduction benefits, and dynamically optimizes and publishes energy trading prices. The lower-layer park agent optimization model adopts the PPO algorithm to perform distributed optimization scheduling of equipment within the park based on the received energy trading prices, with the goal of minimizing its own total operating costs. The allocation module is configured to: improve the Shapley value benefit allocation method based on carbon emission reduction contribution, and fairly allocate the total alliance benefits generated by the collaborative operation of multiple parks; The iterative training module is configured to: conduct collaborative iterative training of the two-layer game optimization model, and through the feedback loop of price signals and scheduling plans, until the game equilibrium is reached, and output the optimal energy trading price and scheduling strategy for each park.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the multi-park integrated energy system electric carbon synergistic optimization method as described in any one of claims 1-7.

10. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the multi-park integrated energy system electricity-carbon synergistic optimization method as described in any one of claims 1-7.

Images