A differentiated carbon pricing demand side electric carbon coupling scheduling optimization method

Abstract

The application discloses a demand side electricity-carbon coupling scheduling optimization method for differentiated carbon pricing. Interconnected users and power plants are obtained as users, participate in the demand side market, determine the actual transaction electricity of each user and the carbon emission intensity attached to the actual transaction electricity; the carbon emission intensity of a single user is calculated according to a carbon emission responsibility balance model, and the carbon emission responsibility of the single user is obtained; the response layer takes the minimization of the total operation cost of the user in a scheduling period as an optimization target, the supervision layer takes the minimization of the comprehensive cost of the whole society in the scheduling period as an optimization target, the supervision layer is taken as a carbon price maker, the response layer is taken as a response party, the carbon price parameter and the user response result are coupled in a closed loop according to a multi-level coordination optimization model, and the ladder carbon price parameter is obtained through real environment closed loop verification. The carbon emission intensity during the discharge of the energy storage is converted from an uncontrollable variable into a controllable variable, and a flexible transaction precondition is provided for differentiated carbon pricing and collaborative scheduling optimization.

Description

Technical Field

[0001] This invention specifically relates to a demand-side electricity-carbon coupling scheduling optimization method for differentiated carbon pricing. Background Technology

[0002] In existing research on demand-side electricity-carbon coupling dispatch optimization, users first engage in electricity trading to determine the transaction price, and then engage in carbon emission responsibility trading to determine the carbon emission intensity associated with the traded electricity volume, with a two-stage, step-by-step clearing process. Regarding carbon emission cost calculation, existing schemes use a fixed carbon price multiplied by the user's total carbon emission responsibility to determine the carbon emission cost. This means all users face a uniform carbon price, failing to consider differences in carbon emission levels and energy flexibility among different users. Furthermore, existing schemes optimize electricity consumption strategies from the perspective of individual users or homogeneous user groups. The carbon price is given as an exogenous parameter, lacking a mechanism from the perspective of the carbon price setter to coordinate the carbon emission reduction behaviors of multiple heterogeneous users.

[0003] According to the proportional sharing theorem, a user's carbon emission responsibility is proportional to their electricity consumption, and this responsibility cannot be changed if the user's electricity consumption remains constant. This rigid coupling restricts the flexible redistribution of carbon emission responsibility among different flexible users, resulting in the inability to fully utilize the adjustment potential of highly flexible users. This limits the flexibility of demand-side electricity-carbon coupling scheduling optimization and the space for carbon emission reduction. Summary of the Invention

[0004] Purpose of the invention: To provide a demand-side electricity-carbon coupling scheduling optimization method for differentiated carbon pricing, which solves the above-mentioned problems existing in the prior art.

[0005] Technical solution: A demand-side electricity-carbon coupling scheduling optimization method for differentiated carbon pricing, comprising the following steps:

[0006] Get individual internet users and Each power plant acts as a user, and each user is equipped with at least an energy storage system, photovoltaic and load equipment. Users, as producers or consumers, participate in the demand-side market and obtain the clearing price for each user. Based on the electric carbon coupled asynchronous market structure model, the actual transaction volume of each user and the carbon emission intensity associated with the actual transaction volume are determined.

[0007] The actual carbon emissions generated by production-side users are obtained, and the actual carbon emissions are allocated to each demand-side user according to the proportional sharing theorem. During the scheduling cycle, the carbon emission intensity of each user is calculated according to the carbon emission responsibility balance model to obtain the carbon emission intensity of each user.

[0008] When a user is in the energy storage discharge state, the carbon emission intensity of electricity sales is calculated based on the controllable carbon intensity model. The carbon emission intensity corresponding to the energy storage is dynamically updated based on the charging and discharging state of the energy storage system, thus transforming the carbon emission intensity during energy storage discharge from an uncontrollable variable into an adjustable one. Simultaneously, by calculating the similarity between the user's load curve and the photovoltaic output distribution, the time-series load demand is matched with the photovoltaic output to reflect the user's ability to track renewable energy output. The response layer aims to minimize the user's total operating cost within the scheduling cycle, while the regulatory layer aims to minimize the overall social cost within the scheduling cycle. With the regulatory layer acting as the carbon price setter and the response layer acting as the responder, the carbon price parameters and user response results are obtained through a closed-loop coupling based on a multi-level coordinated optimization model. The tiered carbon price parameters are then obtained through closed-loop verification in a real environment.

[0009] Preferably, the calculation process for carbon emission intensity is as follows:

[0010] Users, as producers or consumers, use a continuous bilateral auction mechanism to continuously bid on electricity prices, complete electricity transactions, and obtain the first-stage clearing price.

[0011] The clearing price in the first phase is used as a reference to feed back to users. Based on the clearing price in the first phase, users adjust their trading strategies for comprehensive carbon emission costs. The centralized market clearing rules are adopted, and users declare their own carbon emission intensity and expected trading volume. A linear function is used to calculate the carbon intensity declaration curves for both producers and consumers, so as to obtain the actual trading volume and the carbon emission intensity associated with the actual trading volume for each user.

[0012] Preferably, the calculation process for the carbon emission intensity of a single user is as follows:

[0013] The actual carbon emissions generated by users are obtained. The carbon emission responsibility is proportional to the electricity consumption and is allocated to each user on the demand side according to the proportional sharing theorem. The electricity input, electricity output and the corresponding carbon emission responsibility of an individual user should be balanced within the scheduling cycle. The carbon emission intensity of a user is obtained by the carbon emission intensity of purchased electricity and the carbon emission intensity of energy storage output power.

[0014] When a user switches from the power generation side to the demand side, under the condition that the total carbon emission responsibility remains unchanged, the carbon emission intensity of the user's energy storage discharge is transformed from an uncontrollable variable into a controllable variable. The carbon emission intensity of the user's electricity sales when the energy storage is in a discharge state is calculated, and the carbon emission intensity corresponding to the user's energy storage is dynamically updated. An energy storage threshold is preset, and the carbon emission intensity corresponding to the user's energy storage is dynamically updated within the preset energy storage threshold range, thus transforming the carbon emission intensity of the energy storage discharge from an uncontrollable variable into a controllable variable.

[0015] Preferably, the process for reflecting a user's ability to track renewable energy output is as follows:

[0016] The delay settlement factor is calculated based on the distance between the user load distribution and the photovoltaic output distribution. At the end of the trading day cycle, a delayed allocation is carried out, and the total carbon emission responsibility for the trading day cycle is allocated to each user to obtain the total carbon emission responsibility for each user during the trading day cycle.

[0017] Preferably, the process of dynamically updating the carbon emission intensity corresponding to a user's energy storage is as follows:

[0018] A power fluctuation penalty term is introduced to simulate the impact of frequent charge-discharge switching and high-power ramping on user energy storage systems. Based on the energy storage system's charge-discharge power, power ramping amount, number of charge-discharge switching times, and energy storage operating status, the equivalent power loss and cost of the user's energy storage system during the scheduling cycle are calculated. According to the energy storage carbon emission intensity of the energy storage system at the corresponding time, the equivalent power loss is converted into energy storage carbon emission responsibility. Then, time-varying weights and an exponential loss function are introduced to simulate the marginal utility threshold characteristics of users' load reduction, quantify the critical threshold effect of utility loss caused by user electricity consumption reduction, and obtain user points based on the user's load reduction amount, reduction period, low-carbon response contribution, and utility loss degree. An inverse proportional decay model is used as the marginal price. Based on the logarithmic growth characteristics of the incentive compensation corresponding to the user points, the physical boundary of incentive funds is calculated. A carbon responsibility deduction coefficient is constructed based on the user points, and the user points are converted into deductible carbon emission responsibility.

[0019] Preferably, a tiered carbon price is calculated based on a smooth tiered pricing model using the Sigmoid function, so that the carbon price changes smoothly with the increase in emissions, thus obtaining the user's carbon emission responsibility after deferred settlement.

[0020] Preferably, the process for establishing the cost minimization objective for each user is as follows:

[0021] The optimization objective is to minimize the total operating cost within the scheduling cycle. The carbon emission cost within the scheduling cycle is calculated based on the smooth tiered pricing model. The total operating cost includes equipment depreciation, comfort loss, electricity purchase and sale costs, and incentive compensation.

[0022] The process of establishing the regulatory authorities' goal of maximizing social welfare is as follows:

[0023] Within the scheduling cycle, the upper-level optimization objective is to minimize the total social cost, based on the objective function. The user-guided response layer optimizes energy storage charging and discharging strategies and flexible load reduction strategies to achieve dynamic optimization of electricity decarbonization and energy storage carbon emission intensity. The objective function is... This includes the cost of electricity purchased from outside the power grid, the comprehensive operating cost of all users, and the environmental costs of carbon emissions from thermal power plants.

[0024] Preferably, the closed-loop coupling process of obtaining carbon price parameters and user response results based on the multi-level coordinated optimization model is as follows:

[0025] With regulators as the carbon price setters and users as the responders, a multi-level coordinated optimization model is used to calculate the regulators' tiered carbon price parameters and the users' responses, forming a closed-loop coupling through the tiered carbon price parameters and user response results.

[0026] The preferred real-world closed-loop verification process is as follows:

[0027] In the first step, a multi-agent reinforcement learning network was trained offline under a randomly initialized tiered carbon price parameter environment, and the network parameters were frozen. Next, a large-scale carbon price parameter sample was generated using a sampling algorithm and input into the underlying network for forward inference to obtain feature data pairs. This enabled the training of a lightweight agent model that reflects the nonlinear mapping relationship between carbon price parameters and total social cost. Finally, in the online optimization stage, minimizing total social cost was used as the evaluation objective. A high-frequency iterative search was performed directly on the lightweight agent model using a heuristic algorithm. After closed-loop verification in a real environment, the optimal tiered carbon price parameter was issued as the final pricing strategy.

[0028] Beneficial effects: This invention relates to a demand-side electricity-carbon coupling scheduling optimization method for differentiated carbon pricing. By proposing a controllable carbon emission intensity model, the carbon emission intensity during energy storage discharge is transformed from an uncontrollable variable into a controllable variable. This breaks through the rigid coupling constraint of the proportional sharing theorem on carbon emission responsibility and electricity consumption, and realizes the flexible redistribution of carbon emission responsibility among different users. This provides a prerequisite for flexible trading for differentiated carbon pricing and collaborative scheduling optimization.

[0029] By introducing a tiered carbon pricing model to replace the traditional fixed carbon price, differentiated carbon price parameters can be formulated for different types of users based on their carbon emission responsibilities and market clearing results. Tiered carbon pricing imposes stronger emission reduction incentives on high-emission users while maintaining a lower carbon cost burden for low-emission users. This effectively reduces the total carbon emissions of the system while safeguarding social welfare and fully releasing the carbon reduction potential of different flexible users on the demand side.

[0030] By establishing a collaborative optimization framework for carbon pricing and user scheduling, differentiated carbon pricing and user energy consumption strategy optimization in electricity-carbon coupled scheduling are incorporated into a unified model. The carbon pricing authority sets tiered carbon pricing parameters with the goal of maximizing social welfare. Various users independently optimize their own operating strategies based on the given carbon price and feed back the response results to the carbon pricing authority, thus achieving coordination between carbon emission reduction and economic benefits. Attached Figure Description

[0031] Figure 1 This is a system block diagram of the present invention;

[0032] Figure 2 This is a comparison chart of the uniform carbon price and differentiated carbon price parameters of the present invention;

[0033] Figure 3 This invention relates to the user-side hourly load response and energy storage operation. Detailed Implementation

[0034] like Figure 1 As shown, this invention provides a technical solution: a demand-side electricity-carbon coupling scheduling optimization method for differentiated carbon pricing, comprising the following steps:

[0035] Get individual internet users and Each power plant acts as a user, and each user is equipped with at least an energy storage system, photovoltaic and load equipment. Users, as producers or consumers, participate in the demand-side market and obtain the clearing price for each user. Based on the electric carbon coupled asynchronous market structure model, the actual transaction volume of each user and the carbon emission intensity associated with the actual transaction volume are determined.

[0036] The actual carbon emissions generated by production-side users are obtained, and the actual carbon emissions are allocated to demand-side users according to the proportional sharing theorem. Within the scheduling cycle, the carbon emission intensity of each user is calculated based on the carbon emission responsibility balance model. The calculation process for carbon emission intensity is as follows:

[0037] Users, as producers or consumers, use a continuous bilateral auction mechanism to continuously bid on electricity prices, complete electricity transactions, and obtain the first-stage clearing price.

[0038] The clearing price in the first phase will be used as a reference to provide feedback to users. Based on this price, users will adjust their overall carbon emission cost trading strategies. Using the centralized market clearing rules, users will declare their own carbon emission intensity and desired trading volume. A linear function will be used to calculate the carbon intensity declaration curves for both producers and consumers. The calculation formula for these curves is as follows:

[0039] ;

[0040] In the formula: This indicates that the producer or consumer is at The carbon emission intensity corresponding to the transaction volume reported at any time; This indicates the minimum acceptable level of carbon emission intensity for producers or consumers. The reporting slope representing carbon emission intensity; It represents the amount of electricity traded; through centralized market clearing rules, the actual amount of electricity traded by each user and the carbon emission intensity associated with that amount of electricity traded are obtained.

[0041] When a user is in the energy storage discharge phase, the carbon emission intensity of the user during energy storage discharge is transformed from an uncontrollable variable into a controllable variable. Based on the controllable carbon intensity model, the carbon emission intensity when the user sells electricity and is in the discharge state is calculated. The calculation process for the carbon emission intensity of a single user is as follows:

[0042] The actual carbon emissions generated by users are obtained, and the carbon emission responsibility is proportional to electricity consumption. Based on the proportional sharing theorem, these emissions are allocated to each user on the demand side. The electricity input, electricity output, and corresponding carbon emission responsibility of each individual user should remain balanced within the scheduling cycle. The physical constraint for maintaining this balance within the scheduling cycle is calculated using the following formula:

[0043] ;

[0044] In the formula: express Real-time electricity purchase volume; express Electricity sales volume at any given time; express The charge and discharge levels of the energy storage system at any given time; express Solar power output at all times; express The amount of electricity consumed at any given time; express Carbon emission intensity of electricity purchase at any given moment; express Carbon emission intensity of energy storage systems at any given time; express The carbon emission intensity of photovoltaic power output at any given time, among which, ; express Carbon emission intensity of electricity sales at any given time; Carbon emission intensity of electricity load;

[0045] Therefore ;

[0046] The user's carbon intensity is obtained by combining the carbon intensity of purchased electricity with the carbon intensity of energy storage output power.

[0047] When a user transitions from the power generation side to the demand side, under the condition that the total carbon emission responsibility remains unchanged, the carbon emission intensity of the user's energy storage during discharge changes from an uncontrollable variable to a controllable variable. The carbon emission intensity of the user's electricity sales is calculated when the energy storage is in a discharging state, and the carbon emission intensity corresponding to the user's energy storage is dynamically updated. The formula for calculating the carbon emission intensity of electricity sales is as follows:

[0048] ;

[0049] A preset energy storage threshold is set, and the carbon emission intensity of electricity sales meets the preset energy storage threshold constraint, i.e. The carbon emission intensity corresponding to the user's energy storage is dynamically updated within a preset energy storage threshold range, transforming the carbon emission intensity during energy storage discharge from an uncontrollable variable into a controllable one. The dynamic update calculation formula for the carbon emission intensity corresponding to energy storage is as follows:

[0050] ;

[0051] In the formula: express The carbon emission intensity corresponding to the amount of electricity stored in a real-time energy storage system; Indicates the rated capacity of the energy storage system; express The state of charge of the energy storage system at all times; express The carbon emission intensity corresponding to the amount of electricity stored in a real-time energy storage system; express The energy stored in the energy storage system at any time express The charging and discharging capacity of the energy storage system at all times; express The carbon emission intensity corresponding to the amount of electricity charged and discharged by the energy storage system at any given time.

[0052] By comparing the similarity between user load curves and photovoltaic (PV) output distribution, the time-series load is matched with PV power to reflect the user's ability to track renewable energy output. The process of reflecting the user's ability to track renewable energy output is as follows:

[0053] The delay settlement factor is calculated based on the distance between the user load distribution and the photovoltaic output distribution. The formula for calculating the delayed settlement coefficient is as follows:

[0054] ;

[0055] In the formula: Indicates the delayed settlement coefficient; The cumulative distribution function representing the daily load of users; This represents the cumulative distribution function of photovoltaic power generation. Represents a time integral infinitesimal; Indicates the scheduling period The absolute value of the difference between the cumulative distribution function of user daily load and the cumulative distribution function of photovoltaic daily power output is integrated to characterize the cumulative deviation between the user load distribution and the photovoltaic power output distribution.

[0056] The formula for calculating the cumulative distribution function of daily user load is as follows:

[0057] ;

[0058] In the formula: Represents the discrete-time index within the scheduling period; Indicates the user's position in the first month. Load power for each time period; This represents the total number of discrete time slots within the scheduling period; Indicates time;

[0059] At the end of the trading day period, a delayed allocation is performed, distributing the total carbon emission responsibility for the trading day period to each user, resulting in each user's total carbon emission responsibility for the trading day period. The formula for calculating a user's total carbon emission responsibility for the trading day period is as follows:

[0060] ;

[0061] In the formula: Indicates the allocation adjustment coefficient; This represents the total net inflow of basic carbon emissions throughout the day.

[0062] The process of dynamically updating the carbon emission intensity corresponding to a user's energy storage is as follows:

[0063] A power fluctuation penalty term is introduced to simulate the impact of frequent charge-discharge switching and high-power ramping on the user's energy storage system. Based on the mathematical model of the energy storage equipment, calculations are performed to obtain the power loss and cost of the user's energy storage system. Then, according to the carbon emission intensity of the energy storage system at the corresponding time, the power loss is converted into the carbon emission responsibility for energy storage losses. The calculation formula is as follows:

[0064] ;

[0065] In the formula: Indicates the basic aging coefficient; Indicates the cycle life pressure coefficient; express The charge and discharge levels of the energy storage system at any given time;

[0066] By introducing time-varying weights and an exponential loss function, the marginal utility threshold characteristics of users' load reduction are simulated, quantifying the critical threshold effect of utility loss caused by users' electricity consumption reduction. The user integral is obtained based on the user's load reduction amount, reduction period, low-carbon response contribution, and degree of utility loss. The calculation formula for the user integral is as follows:

[0067] ;

[0068] In the formula: This represents the sensitivity factor, which determines how steeply the loss function rises with the amount of reduction; Indicates the basic price coefficient; Indicates time-varying weights; Indicate The exponential marginal utility loss term corresponding to the load reduction at any given moment; Indicate The cost of dynamic comfort loss for users due to load reduction at any given moment; Indicate The exponential marginal utility loss term corresponding to the load reduction at any given moment;

[0069] Using an inverse proportional decay model as the marginal price, and based on the logarithmic growth characteristic of the incentive compensation corresponding to user points, the physical boundary of the incentive funds is calculated. The carbon responsibility deduction coefficient is constructed based on the user points, and the calculation formula for converting user points into the physical boundary of deductible carbon emission responsibility incentive funds is as follows:

[0070] ;

[0071] In the formula: Indicates the initial incentive rate. Indicates the saturation coefficient; Indicates that the user is The amount of load reduction at any given time; This indicates the physical boundaries of the incentive funds.

[0072] Based on the smooth tiered pricing model using the Sigmoid function, a tiered carbon price is calculated, ensuring a smooth dynamic update of the carbon price as emissions increase, thus yielding the user's carbon emission responsibility after deferred settlement. The formula for calculating the tiered carbon price is as follows:

[0073] ;

[0074] In the formula: Indicates the base carbon price (starting price); This indicates the user's cumulative carbon emissions; This indicates the first trigger for a price increase. Step threshold; This indicates the price increase range for that tier; This represents the smoothing factor hyperparameter; Indicates a tiered carbon price; Represents the natural constant; This represents the total number of tiers in the tiered carbon pricing model; Indicates the ladder number.

[0075] The response layer aims to minimize the total operating cost of users within the scheduling period, while the regulatory layer aims to minimize the overall cost to society within the scheduling period. With the regulatory layer acting as the carbon price setter and the response layer as the responder, the process for establishing the cost minimization objective for each user is as follows:

[0076] Minimizing the total operating cost within the scheduling period is the optimization objective. The carbon emission cost within the scheduling period is calculated based on a smooth tiered pricing model, using the following formula:

[0077] ;

[0078] In the formula: This represents the net operating cost after combining all costs and revenues throughout the entire scheduling cycle for a single user. This indicates the loss cost of the user's energy storage system; express The cost of dynamic comfort loss for users due to load reduction at any given moment; express The electricity cost incurred by users when purchasing electricity at any given time; express The electricity revenue that users receive from selling electricity at any given time; express The amount of saturation incentive compensation received by users at any given moment; This represents a smoothed step carbon valence function; This indicates the amount of carbon emission responsibility ultimately borne by the user after the delayed settlement; the total operating cost includes equipment depreciation, loss of comfort, electricity purchase and sale costs, and incentive compensation.

[0079] The process of establishing the regulatory authorities' goal of maximizing social welfare is as follows:

[0080] Within the scheduling cycle, the upper-level optimization objective is to minimize the total social cost, based on the objective function. The system guides users in the response layer to optimize energy storage charging and discharging strategies and flexible load reduction strategies, thereby achieving dynamic optimization of low-carbon electricity consumption and energy storage carbon emission intensity. The calculation formula is as follows:

[0081] ;

[0082] In the formula: express The marginal purchase price or marginal generation cost of the large power grid at any given time; express Net power exchange between the timekeeping system and the main power grid; This represents the sum of the total operating costs for all users within the scheduling period; This represents the social and environmental cost coefficient corresponding to a unit of carbon emissions; This represents the cumulative carbon emissions corresponding to the system's net electricity purchases from the main grid during the scheduling period; where, the objective function is... This includes the cost of electricity purchased from outside the power grid, the comprehensive operating cost of all users, and the environmental costs of carbon emissions from thermal power plants.

[0083] Then, based on the multi-level coordinated optimization model, the carbon price parameter and the user response result are obtained in a closed loop coupling. The process of obtaining the carbon price parameter and the user response result in a closed loop coupling based on the multi-level coordinated optimization model is as follows:

[0084] With regulators as the carbon price setters and users as the responders, the tiered carbon price parameters of regulators and user responses are calculated using a multi-level coordinated optimization model. The calculation formula is as follows:

[0085] ;

[0086] ;

[0087] In the formula: This refers to the set of tiered carbon pricing parameters issued by regulators to each user. The optimization variables are minimized, including the set of tiered carbon price parameters distributed by regulators to each user. For upper-level decision variables; Indicates the first The optimal response strategy for a user under a given tiered carbon price parameter; Let represent the decision variables of the response layer, where the decision variables of the response layer are the first _____. Individual user operation strategy Objective function Depends on the optimal response results of all users The optimal response of each user depends on the tiered carbon price parameters issued by the regulator. A closed-loop coupling is formed between the tiered carbon price parameters and the user response results. To overcome the convergence difficulties caused by frequent iterations of multiple agents in the regulator and response layers, the underlying multi-agent reinforcement learning network is trained offline under a randomly initialized tiered carbon price parameter environment, and the network parameters are frozen. Secondly, a large-scale carbon price parameter sample is generated through a sampling algorithm and input into the underlying network for forward inference to obtain feature data pairs, thereby training a lightweight agent model that reflects the nonlinear mapping relationship between carbon price parameters and total social cost. Finally, in the online optimization stage, the minimization of total social cost is used as the evaluation objective. A high-frequency iterative search is performed directly on the lightweight agent model using a heuristic algorithm. After closed-loop verification in a real environment, the optimal tiered carbon price parameter is issued and executed as the final pricing strategy.

[0088] To verify the feasibility of the demand-side electricity-carbon coupling scheduling optimization method for differentiated carbon pricing described in this invention, a simulation system was constructed comprising four interconnected demand-side users. Each user is equipped with load, photovoltaic, and energy storage devices, and can make flexible load reduction, energy storage charging and discharging, and electricity trading decisions based on electricity price signals, carbon price signals, and their own energy storage status. The simulation compared various schemes, including an electricity benchmark, a fixed carbon price, a default tiered carbon price, a unified optimized tiered carbon price, a differentiated tiered carbon price, and the complete mechanism of this invention.

[0089] Simulation results are shown in Table 1. Under the electricity baseline scheme, the system's average daily positive carbon emission liability is 10154.70 kg. After adopting the complete mechanism of this invention, the average daily positive carbon emission liability is reduced to 8607.66 kg, the average daily net carbon emission liability is reduced to 7516.87 kg, and the average daily low-carbon credits exported reach 1090.78 kg. Compared with the electricity baseline scheme, the complete mechanism of this invention reduces the daily carbon emission liability by 15.23% and improves social welfare by 14.02%, indicating that this method can maintain good economic efficiency while reducing demand-side carbon emission liability.

[0090] Table 1. Comparison of the complete mechanism of this invention with the power benchmark scheme.

[0091]

[0092] To reflect the regulatory authorities' optimization of carbon price parameters, a further comparison was made between the default tiered carbon price, the unified optimized tiered carbon price, and the differentiated optimized tiered carbon price. As shown in Table 2, the differentiated optimized tiered carbon price, compared to the default tiered carbon price, reduced daily carbon emission liability by 1.49%, increased total rewards by 8.33%, and improved social welfare by 1.32%. This result indicates that setting different base carbon prices, tiered thresholds, and tiered increments for different users can more fully reflect the differences in users' carbon emission levels and adjustment capabilities.

[0093] Table 2. Effects of Regulatory Parameter Optimization

[0094]

[0095] like Figure 2 As shown, to improve the efficiency of carbon price parameter optimization at the regulatory level, this embodiment uses a lightweight surrogate model to fit the nonlinear mapping relationship between carbon price parameters and system operation results. As shown in Table 3, the prediction determination coefficients of the surrogate model for social welfare, daily carbon emission responsibility, and total reward are 0.973, 0.975, and 0.978, respectively, indicating that the surrogate model can well reflect the impact of carbon price parameter changes on system operation results and can be used for rapid search of upper-level carbon price parameters.

[0096] Table 3 Prediction Accuracy of the Proxy Model

[0097]

[0098] After obtaining the optimized differentiated tiered carbon pricing parameters, they are distributed to the user response layer. Users make scheduling decisions based on hourly electricity prices, carbon prices, load demand, and energy storage status. During periods of high load or strong carbon price constraints, they reduce net electricity demand through flexible load shedding and energy storage charging and discharging, thereby achieving carbon emission reduction and operating cost optimization. The user-side hourly response process is as follows: Figure 3 As shown.

[0099] In summary, this embodiment forms a closed-loop optimization process of "carbon price parameter setting—user-side response—operational result feedback—real-world verification". Simulation data shows that this invention can guide users to adjust their energy consumption behavior through differentiated tiered carbon pricing, reduce demand-side carbon emission responsibility while ensuring the feasibility of the method, and unleash the low-carbon adjustment potential of different users.

[0100] The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited to the specific details of the above embodiments. Within the scope of the technical concept of the present invention, various equivalent transformations can be made to the technical solutions of the present invention, and these equivalent transformations all fall within the protection scope of the present invention.

Claims

1. A demand-side electricity-carbon coupling scheduling optimization method for differentiated carbon pricing, characterized in that, Includes the following steps: Get individual internet users and Each power plant acts as a user, and each user is equipped with at least an energy storage system, photovoltaic and load equipment. Users, as producers or consumers, participate in the demand-side market and obtain the clearing price for each user. Based on the electric carbon coupled asynchronous market structure model, the actual transaction volume of each user and the carbon emission intensity associated with the actual transaction volume are determined. The actual carbon emissions generated by production-side users are obtained, and the actual carbon emissions are allocated to each demand-side user according to the proportional sharing theorem. During the scheduling cycle, the carbon emission intensity of each user is calculated according to the carbon emission responsibility balance model to obtain the carbon emission intensity of each user. When a user is in the energy storage discharge state, the carbon emission intensity of electricity sales is calculated based on the controllable carbon intensity model, where the user is selling electricity while the energy storage is in the discharge state. The carbon emission intensity corresponding to the energy storage is dynamically updated based on the charging and discharging state of the energy storage system, thus transforming the carbon emission intensity during energy storage discharge from an uncontrollable variable into an adjustable one. Simultaneously, by calculating the similarity between the user's load curve and the photovoltaic output distribution, the time-series load demand is matched with the photovoltaic output to reflect the user's ability to track renewable energy output. The response layer aims to minimize the user's total operating cost within the dispatch cycle, while the regulatory layer aims to minimize the overall social cost within the dispatch cycle. With the regulatory layer acting as the carbon price setter and the response layer acting as the responder, a closed-loop coupling of carbon price parameters and user response results is obtained based on a multi-level coordinated optimization model. This closed-loop verification in a real environment yields the tiered carbon price parameters.

2. The demand-side electricity-carbon coupling scheduling optimization method for differentiated carbon pricing according to claim 1, characterized in that, The calculation process for carbon emission intensity is as follows: Users, as producers or consumers, use a continuous bilateral auction mechanism to continuously bid on electricity prices, complete electricity transactions, and obtain the first-stage clearing price. The clearing price in the first phase is used as a reference to feed back to users. Based on the clearing price in the first phase, users adjust their trading strategies for comprehensive carbon emission costs. The centralized market clearing rules are adopted, and users declare their own carbon emission intensity and expected trading volume. A linear function is used to calculate the carbon intensity declaration curves for both producers and consumers, so as to obtain the actual trading volume and the carbon emission intensity associated with the actual trading volume for each user.

3. The demand-side electricity-carbon coupling scheduling optimization method for differentiated carbon pricing according to claim 1, characterized in that, The calculation process for a single user's carbon emission intensity is as follows: The actual carbon emissions generated by users are obtained. The carbon emission responsibility is proportional to the electricity consumption and is allocated to each user on the demand side according to the proportional sharing theorem. The electricity input, electricity output and the corresponding carbon emission responsibility of an individual user should be balanced within the scheduling cycle. The carbon emission intensity of a user is obtained by the carbon emission intensity of purchased electricity and the carbon emission intensity of energy storage output power. When a user switches from the power generation side to the demand side, under the condition that the total carbon emission responsibility remains unchanged, the carbon emission intensity of the user's energy storage discharge is transformed from an uncontrollable variable into a controllable variable. The carbon emission intensity of the user's electricity sales when the energy storage is in a discharge state is calculated, and the carbon emission intensity corresponding to the user's energy storage is dynamically updated. An energy storage threshold is preset, and the carbon emission intensity corresponding to the user's energy storage is dynamically updated within the preset energy storage threshold range, thus transforming the carbon emission intensity of the energy storage discharge from an uncontrollable variable into a controllable variable.

4. The demand-side electricity-carbon coupling scheduling optimization method for differentiated carbon pricing according to claim 3, characterized in that, The process for reflecting users' ability to track renewable energy output is as follows: The delay settlement factor is calculated based on the distance between the user load distribution and the photovoltaic output distribution. At the end of the trading day cycle, a delayed allocation is carried out, and the total carbon emission responsibility for the trading day cycle is allocated to each user to obtain the total carbon emission responsibility for each user during the trading day cycle.

5. The demand-side electricity-carbon coupling scheduling optimization method for differentiated carbon pricing according to claim 3, characterized in that, The process of dynamically updating the carbon emission intensity corresponding to a user's energy storage is as follows: A power fluctuation penalty term is introduced to simulate the impact of frequent charge-discharge switching and high-power ramping on user energy storage systems. Based on the energy storage system's charge-discharge power, power ramping amount, number of charge-discharge switching times, and energy storage operating status, the equivalent power loss and cost of the user's energy storage system during the scheduling cycle are calculated. According to the energy storage carbon emission intensity of the energy storage system at the corresponding time, the equivalent power loss is converted into energy storage carbon emission responsibility. Then, time-varying weights and an exponential loss function are introduced to simulate the marginal utility threshold characteristics of users' load reduction, quantify the critical threshold effect of utility loss caused by user electricity consumption reduction, and obtain user points based on the user's load reduction amount, reduction period, low-carbon response contribution, and utility loss degree. An inverse proportional decay model is used as the marginal price. Based on the logarithmic growth characteristics of the incentive compensation corresponding to the user points, the physical boundary of incentive funds is calculated. A carbon responsibility deduction coefficient is constructed based on the user points, and the user points are converted into deductible carbon emission responsibility.

6. The demand-side electricity-carbon coupling scheduling optimization method for differentiated carbon pricing according to claim 3, characterized in that, Based on the smooth tiered pricing model of the Sigmoid function, a tiered carbon price is calculated, which makes the dynamic update of the carbon price as emissions increase smoothly transition, and obtains the user's carbon emission responsibility after deferred settlement.

7. The demand-side electricity-carbon coupling scheduling optimization method for differentiated carbon pricing according to claim 5, characterized in that, The process for establishing the goal of minimizing the cost per user is as follows: The optimization objective is to minimize the total operating cost within the scheduling cycle. The carbon emission cost within the scheduling cycle is calculated based on the smooth tiered pricing model. The total operating cost includes equipment depreciation, comfort loss, electricity purchase and sale costs, and incentive compensation. The process of establishing the regulatory authorities' goal of maximizing social welfare is as follows: Within the scheduling cycle, the upper-level optimization objective is to minimize the total social cost, based on the objective function. The user-guided response layer optimizes energy storage charging and discharging strategies and flexible load reduction strategies to achieve dynamic optimization of electricity decarbonization and energy storage carbon emission intensity. The objective function is... This includes the cost of electricity purchased from outside the power grid, the comprehensive operating cost of all users, and the environmental costs of carbon emissions from thermal power plants.

8. The demand-side electricity-carbon coupling scheduling optimization method for differentiated carbon pricing according to claim 6, characterized in that, The closed-loop coupling process between carbon price parameters and user response results obtained from the multi-level coordinated optimization model is as follows: With regulators as the carbon price setters and users as the responders, a multi-level coordinated optimization model is used to calculate the regulators' tiered carbon price parameters and the users' responses, forming a closed-loop coupling through the tiered carbon price parameters and user response results.

9. The demand-side electricity-carbon coupling scheduling optimization method for differentiated carbon pricing according to claim 7, characterized in that, The real-world closed-loop verification process is as follows: In the first step, a multi-agent reinforcement learning network was trained offline under a randomly initialized tiered carbon price parameter environment, and the network parameters were frozen. Next, a large-scale carbon price parameter sample was generated using a sampling algorithm and input into the underlying network for forward inference to obtain feature data pairs. This enabled the training of a lightweight agent model that reflects the nonlinear mapping relationship between carbon price parameters and total social cost. Finally, in the online optimization stage, minimizing total social cost was used as the evaluation objective. A high-frequency iterative search was performed directly on the lightweight agent model using a heuristic algorithm. After closed-loop verification in a real environment, the optimal tiered carbon price parameter was issued as the final pricing strategy.

Images