A low-carbon optimal dispatching method for active power distribution network considering marginal carbon potential demand response of nodes

Abstract

The present application relates to a kind of active distribution network low-carbon optimization scheduling method considering node marginal carbon potential demand response, comprising: S1, the double-layer collaborative low-carbon scheduling framework of distribution network and user is established, and node marginal carbon potential fast calculation model is established;S2, the lowest upper layer system comprehensive operation cost is calculated by upper layer distribution network optimization scheduling model;S3, the marginal relationship between the unit increment of node load and the increment of system overall carbon emission is accurately quantified using node marginal carbon potential fast calculation model;S4, the maximum user side comprehensive benefit is calculated by lower layer low-carbon demand response model considering DLMCI;S5, double-layer optimization model is established, and an operable interval containing target optimal operating state is obtained;S6, repeat S2 to S5, output final scheduling result.Compared with prior art, the present application has the advantages of carbon reduction guidance precision, promoting renewable energy consumption, high solving efficiency, strong convergence stability, considering economy and low carbon, reducing system comprehensive operation cost and the like.

Description

Technical Field

[0001] This invention relates to the field of power system automation and low-carbon operation technology, and in particular to an active distribution network low-carbon optimization scheduling method that takes into account the marginal carbon potential demand response of nodes. Background Technology

[0002] The power system's operating mode is rapidly evolving towards a "safe, economical, and low-carbon" model. As the "last mile" of power supply, the distribution network has a large amount of dispatchable flexible load resources on the user side, such as electric vehicles and smart home appliances. Effectively tapping the carbon reduction potential of these resources is a crucial means to achieve green and low-carbon operation of the power grid.

[0003] Existing low-carbon demand response research largely relies on dynamic / average carbon emission factors calculated using carbon emission flow theory as guiding signals. However, existing technologies suffer from the following serious drawbacks: First, the average carbon emission factor is essentially based on the allocation of carbon emission responsibility on the load side, and cannot capture and reflect the marginal changes in the total carbon emissions of the system caused by adjustments in users' micro-level electricity consumption behavior. In other words, it cannot identify whether increases or decreases in load have triggered the start-up or shutdown of high-emission standby units or the curtailment of renewable energy.

[0004] A search of Chinese Patent Publication No. CN118380993A reveals a distribution network scheduling method based on a two-way feedback mechanism of carbon potential load data, aimed at achieving coordinated low-carbon and economical operation at the distribution network level. This method, based on carbon emission flow theory, constructs load and energy storage carbon emission flow models at the distribution network level, establishes a demand response priority function to assess the demand response capabilities of various loads, and designs corresponding demand response strategies based on this priority. Combining the economic and low-carbon aspects of the distribution network, a two-way feedback mechanism between upper-level carbon potential data and lower-level load data is designed, and a two-layer low-carbon scheduling model for the distribution network is established and solved based on this mechanism. However, this existing patent uses dynamic or average carbon emission factors as guiding signals, which cannot capture the marginal changes in the total carbon emissions of the system due to users' micro-level electricity consumption behavior, leading to a disconnect between the carbon incentive signal and the actual carbon reduction benefits. Furthermore, the iterative solution of the existing patent's two-layer model is prone to oscillations, resulting in slow convergence rates, low operating efficiency, and even difficulty in obtaining the global optimal solution.

[0005] Therefore, how to capture the marginal changes in the total carbon emissions of the system caused by users' micro-level electricity consumption behavior, and how to reduce the oscillation phenomenon caused by the iterative solution of the two-level model, have become problems that need to be solved. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the existing technology and provide an active distribution network low-carbon optimization scheduling method that takes into account the marginal carbon potential demand response of nodes.

[0007] The objective of this invention can be achieved through the following technical solutions: A proactive distribution network low-carbon optimization scheduling method considering nodal marginal carbon potential demand response includes the following steps: Step S1: Establish a two-layer collaborative low-carbon dispatching framework for the distribution network and users, including an upper-layer distribution network optimization dispatching model and a lower-layer low-carbon demand response model considering DLMCI; and establish a fast calculation model for node marginal carbon potential. Step S2: Calculate the lowest overall operating cost of the upper-level system using the upper-level distribution network optimization scheduling model; Step S3: Accurately quantify the DLMCI using the fast calculation model of the node marginal carbon potential; the DLMCI is the marginal relationship between the unit increment of node load and the overall carbon emission increment of the system. Step S4: Calculate the maximum comprehensive user-side benefits using the lower-level low-carbon demand response model that considers DLMCI. Step S5: Establish a two-layer optimization model and obtain an operable range containing the optimal operating state of the target through the bisection mechanism; Step S6: Repeat steps S2 to S5 until the length of the operable interval is less than the preset convergence accuracy, determine that the model has converged, and output the final scheduling result.

[0008] As a preferred technical solution, the lowest overall operating cost of the upper-level system includes the cost of power curtailment. Electricity purchase cost from the upper-level power grid Energy storage costs Gas turbine cost and carbon emission costs .

[0009] As a preferred technical solution, the cost of wind and solar curtailment The specific calculations are as follows: in: and These represent the output power of the wind turbine and the photovoltaic system at time t, respectively. , These represent the maximum output power of the wind turbine and the photovoltaic system at time t, respectively. , These are the penalty cost coefficients for wind and solar power curtailment, respectively. , These are collections of wind turbines and photovoltaic units, respectively. The cost of electricity purchase from the upper-level power grid The specific calculations are as follows: in: Let be the time-of-use electricity price at time t; Let t be the power purchased by the distribution network from the upper-level grid; Energy storage costs The specific calculations are as follows: in: This is the operating cost coefficient for the energy storage system. Let t be the charging power of the energy storage system at time t; Let be the discharge power of the energy storage system at time t; The cost of the gas turbine The specific calculations are as follows: in: This refers to the set of distribution network nodes connected to the gas turbine. , , These are the quadratic, linear, and constant coefficients of the power generation cost of the gas turbine at node i, respectively. Let be the active power output of the gas turbine connected to node i at time t; The carbon emission cost The specific calculations are as follows: in: Carbon emission coefficient; , These represent the carbon emission intensity of the upstream power grid's electricity purchase and the energy storage system at time t, respectively. The carbon emission intensity of the gas turbine connected to node i; The minimum overall operating cost of the upper-level system is calculated as follows: .

[0010] As a preferred technical solution, the upper-level distribution network optimization scheduling model also includes main grid power purchase constraints, distributed unit operation constraints, energy storage charging and discharging constraints, distribution network power flow constraints, and node power balance constraints.

[0011] As a preferred technical solution, the fast calculation model of node marginal carbon potential is based on the economic dispatch model of distribution network with distributed power sources. It uses the Lagrange multiplier method, dual vector and complementary relaxation conditions to establish an equivalent model. It uses the parameter differential perturbation method to derive and establish an analytical expression to accurately quantify the marginal relationship between the unit increment of node load and the overall carbon emission increment of the system. The compact form of the fast calculation model for the nodal marginal carbon potential is expressed as follows: Where A and b are the quadratic coefficient matrix and linear coefficient vector of the power generation cost, respectively, and the coefficient corresponding to the new energy power station is 0; It is the power output vector formed by the power source. E, D, and h are the corresponding coefficient matrices or vectors, and , ; Let be the dual vector of the constraint. This represents the active power output vector of all generating units in the distribution network.

[0012] As a preferred technical solution, based on the Lagrange multiplier method and the Karush-Kuhn-Tucker condition, the compact form of the Lagrange function representing the lowest overall operating cost of the upper-level system is: The Lagrange equation is: The complementary relaxation condition is: in," "Indicates Hadama accumulation; The rate of change in the overall carbon emissions of the system when a node experiences a small change in load is defined as the node marginal carbon potential, expressed by the formula: in, This represents the total carbon emissions of the power distribution network. This represents the active power output vector of all generating units in the distribution network. It is by , and The power output vector formed by the power source ; This is the active power output vector of the higher-level system. This represents the active power output vector of all new energy power stations across all time periods. Let be the carbon emission factor of the i-th unit. Let i be the active power output vector of the i-th generating unit in the distribution network. The marginal carbon emission factor of the upper-level power grid; For the above The derivation is as follows: in ; For the aforementioned incremental and Using the parameter differential perturbation method to The solution is performed, and the specific derivation is as follows: Differentiating the Lagrange equation yields: in Therefore, we can obtain From the above The derivation yields: We can obtain: Finally, we can seek the aforementioned... for: In conclusion: .

[0013] As a preferred technical solution, the maximum comprehensive benefits on the user side include: user energy efficiency, energy purchase cost, and low-carbon response incentives.

[0014] As a preferred technical solution, the user energy efficiency can be expressed as: : in: for t Immediate demand response afterload; and The electricity efficiency coefficient for users; Energy purchase cost It can be represented as: in: Developed for the upper-level system t The transaction price at any given moment; The low-carbon response incentive It can be represented as in: for t Time Node i Load after demand response The objective function for maximizing overall user-side benefits is expressed as follows: .

[0015] As a preferred technical solution, the maximum comprehensive user-side benefit is subject to constraints, including interruptible load constraints, transferable load constraints, and demand response power balance constraints. The interruptible load constraints include interruptible capacity, number of interruptions, maximum allowable interruption duration, and minimum time interval between two interruptions; in: for t Time Node i Reduced active power load; The maximum interruption amount for interruptible loads; To indicate t Time Node i A 0-1 variable indicating whether the load is reduced; For nodes i The maximum number of interruptible load interruptions; The maximum interruption duration for interruptible loads; The minimum interruption interval for interruptible loads; The transferable load comfort includes, in: and They are respectively t Time Node i Active load transferred in and out; and These are the maximum amount of load that can be transferred in and the maximum amount of load that can be transferred out, respectively. for t Time Node i 0-1 variables representing transferable loads; The demand response power balance constraint consists of interruptible load constraints and transferable load constraints. The demand response power balance constraint is as follows: .

[0016] As a preferred technical solution, steps S5 and S6 specifically include: Initialize the operable range of the target's optimal operating state; After each double-layer iteration, based on the comparison between the upper-layer objective function value and the midpoint objective value of the interval, if the current value is better than the midpoint value, the corresponding sub-interval is retained; otherwise, it is discarded. By continuously narrowing the search interval, the oscillations of the DLMCI signal and load response are forcibly suppressed until the interval length is less than the preset convergence accuracy, at which point the model is determined to have converged. After convergence, the final scheduling results are output: the upper-level main grid power purchase plan, energy storage charging and discharging strategy, and gas turbine output curve; the lower-level interruptible and transferable load response curves; and the carbon emissions and comprehensive operating cost indicators of the entire system.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. More precise carbon reduction guidance avoids ineffective carbon reduction. It can accurately reflect the marginal impact of users' electricity consumption behavior on the system's carbon emissions, guiding reductions where increased load would increase carbon emissions and guiding increases where increased load could absorb green electricity, resulting in significantly higher carbon reduction efficiency than traditional methods.

[0018] Existing methods often employ average carbon emission factors, which fail to identify whether load changes trigger the start-up or shutdown of high-emission units or the curtailment of renewable energy. This invention constructs a rapid calculation model for the marginal carbon potential of nodes adapted to the radial topology of the distribution network. By deriving the analytical relationship between the unit increment of node load and the total carbon emission increment of the system through the parametric differential perturbation method, it quantifies the real marginal emission reduction benefits of micro-level electricity consumption behavior from a mechanistic perspective, upgrading from average allocation to marginal source tracing.

[0019] 2. Fully explore the carbon reduction potential of the distribution network on the user side and promote the consumption of renewable energy. It effectively mobilized the adjustment capacity of flexible loads such as electric vehicles and smart home appliances, significantly reduced the curtailment rate of wind and solar power, and improved the local consumption level of distributed renewable energy.

[0020] Traditional demand response relies heavily on time-of-use pricing signals, with users adjusting load solely for economic reasons and lacking a low-carbon orientation. This invention introduces a low-carbon response incentive mechanism based on DLMCI into the lower-level user model, superimposing carbon price signals with electricity price signals. Users can receive additional low-carbon incentives for consuming more electricity during low DLMCI periods, proactively shifting their load to periods with high renewable energy availability, thus achieving bidirectional coordination between source-driven load and load-driven source.

[0021] 3. High solution efficiency, strong convergence stability, and adaptable to large-scale distribution network dispatching. It solves the problems of iterative oscillation, slow convergence or even non-convergence caused by strong source-load coupling in the two-layer model, significantly improves the solution speed, and meets the timeliness requirements of day-ahead / intra-day scheduling of active distribution networks.

[0022] Traditional two-level iterations employ a direct feedback method, where the interplay between the carbon potential signal and load response can easily lead to oscillations. This invention innovatively introduces a bisection convergence acceleration mechanism, maintaining an operational interval containing the optimal solution for distribution network operators. Each iteration dynamically narrows this interval by comparing the objective function value, forcibly smoothing fluctuations in the DLMCI signal and load response, thus ensuring the algorithm's robustness and solution efficiency in scenarios with large-scale nodes and multiple flexible loads.

[0023] 4. Balancing economic efficiency and low carbon footprint to reduce overall system operating costs. While achieving deep carbon reduction, it effectively controlled the overall operating cost of the power distribution network, realizing multi-objective optimization of safety, economy, and low carbon.

[0024] The upper-level model aims to minimize overall operating costs, taking into account the costs of wind and solar curtailment, electricity purchase, energy storage, gas turbine, and carbon emission costs. The lower-level model aims to maximize overall user benefits, and through reasonable carbon incentive pricing design, it ensures that users' carbon reduction benefits match the system's carbon reduction costs, avoiding the additional economic burden caused by excessive incentives, and achieving a win-win situation for both social benefits and user interests.

[0025] 5. Highly adaptable and can be directly applied to actual power distribution network projects. The model closely matches the actual physical characteristics of the distribution network and can be deployed without large-scale modifications to existing power grid hardware.

[0026] This invention is specifically designed to model radial topology, high proportion of distributed power source access, and Distflow branch power flow characteristics of distribution networks. It employs second-order cone relaxation technology to convexify nonlinear power flow constraints, ensuring both the physical realism of the model and the convenience of mathematical solution. It can be directly applied to low-carbon dispatching scenarios in IEEE standard distribution networks or actual urban distribution networks. Attached Figure Description

[0027] Figure 1 A schematic diagram of a low-carbon demand response mechanism based on marginal carbon potential; Figure 2 Flowchart for solving the model; Figure 3 This is a topology diagram of the IEEE 33-node network. Figure 4 A graph showing renewable energy and load output; Figure 5 A full-day DLMCF heatmap; Figure 6 Power balance diagram for scenario 3; Figure 7 This is to reduce the load demand response result; Figure 8This represents the results of the transferable load demand response; Figure 9 Power balance diagram for scenario 1; Figure 10 This is the power balance diagram for scenario 2. Detailed Implementation

[0028] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0029] Example 1: like Figure 1 The diagram shows a low-carbon demand response mechanism based on marginal carbon potential. Step S1: Establish a two-layer collaborative low-carbon dispatching framework for the distribution network and users, including an upper-layer distribution network optimization dispatching model and a lower-layer low-carbon demand response model considering DLMCI; establish a fast calculation model for node marginal carbon potential. Step S2 involves calculating the lowest overall operating cost of the upper-level system using the upper-level distribution network optimization scheduling model. This cost covers wind and solar curtailment costs, upper-level grid power purchase costs, energy storage costs, gas turbine costs, and carbon emission costs. It also incorporates main grid power purchase constraints, distributed generation unit operation constraints, and energy storage charging and discharging constraints. Furthermore, it utilizes second-order cone relaxation techniques to convexly process Distflow branch power flow constraints, ensuring the economic efficiency and security of the distribution network operation. Step S3: Using the fast calculation model of node marginal carbon potential based on the economic dispatch model of distribution network with distributed power sources, an equivalent model is established using the Lagrange multiplier method, dual vectors and complementary relaxation conditions. The analytical expression is derived and established using the parameter differential perturbation method to accurately quantify the marginal relationship between the unit increment of node load and the overall carbon emission increment of the system.

[0030] Step S4: Calculate the maximum comprehensive user-side benefits through the lower-level low-carbon demand response model that considers DLMCI, including user energy efficiency, time-of-use energy purchase costs, and low-carbon response incentive benefits. Establish capacity, time period, and power balance constraints for interruptible loads and transferable loads respectively, guide users to flexibly adjust their energy consumption strategies in response to dynamic carbon potential signals, and deeply explore the carbon reduction potential on the demand side.

[0031] Step S5: Establish a two-layer optimization model to solve the iterative oscillation problem caused by the strong coupling between the lowest comprehensive operating cost of the upper-layer system and the largest comprehensive benefit on the user side through a bisection mechanism. This results in an operable interval that includes the target optimal operating state. The upper and lower limits of the interval are dynamically updated in each iteration to narrow the range, thereby forcibly suppressing the oscillation between the DLMCI signal and the load response, and ensuring the solution efficiency and convergence of the two-layer model.

[0032] Step S6: Repeat steps S2 to S5 until the interval length is less than the preset convergence accuracy. Determine that the model has converged and output the converged upper-level equipment output plan and lower-level load response curve.

[0033] Example 2: like Figure 2 The diagram shown is a flowchart of the model solution process of this invention; 1. Process Start Point and Initialization Start: Process initiated.

[0034] Initialize the number of iterations k=0; Load data, unit output characteristics and grid topology: Input basic data of the distribution network (such as load of each node, output capacity of units (gas turbine, distributed power source), and line / node topology of the grid).

[0035] 2. Upper-level optimization Based on basic data, distribution network operators obtain initial unit output plans and total carbon emissions: The upper layer aims to achieve the lowest overall system cost (including wind and solar curtailment, electricity purchase, energy storage, gas turbines, and carbon emission costs) by solving the initial dispatch scheme of the distribution network (such as unit output, energy storage charging and discharging, and main grid electricity purchase), and calculating the total carbon emissions at this time.

[0036] The DLMCI (Digital Dynamics Model for Calculating Node Marginal Carbon Potential) is then published to the user side. Based on the KKT optimality condition of the upper-level scheduling, the node marginal carbon potential is derived through the parameter differential perturbation method. The DLMCI is then published to the user side as a low-carbon guidance signal.

[0037] 3. Lower-level optimization Users respond to low-carbon demands based on node marginal carbon potential signals: On the user side, aiming to maximize overall benefits, after receiving the DLMCI signal, the flexible load is optimized as follows: Interruptible load: Reduce load during periods of high DLMCI (reduce high carbon emissions); Transferable load: Shifting loads from low DLMCI periods to high DLMCI periods (avoiding high carbon periods).

[0038] Calculate the overall energy efficiency of user consumption, optimize and update energy consumption behavior: The user side solves the demand response model to obtain the updated load curve (such as the adjusted interruptible / transferable load) and feeds the new load data back to the upper layer.

[0039] 4. Convergence Judgment and Iteration Control Has it converged? Check whether the iteration meets the convergence conditions (such as the change of the objective function between the upper and lower layers being less than the threshold, the fluctuation of DLMCI being less than the threshold, the number of iterations reaching the upper limit, etc.).

[0040] If so: Output the final low-carbon optimization scheduling strategy for the power distribution network (including unit output, energy storage, user load adjustment schemes, etc.), and the process ends.

[0041] If not: Execute k=k+1 (increment the iteration count by 1), return to the upper-level optimization stage, and repeat the cycle of "upper-level scheduling → calculate DLMCI → lower-level response → update load".

[0042] 5. Iterative oscillation constrained by the bisection method: To avoid non-convergence due to iterative oscillations between upper and lower layers, a bisection method is introduced to dynamically narrow the optimization interval, forcing rapid convergence of the iterations.

[0043] 6. k=k+1: Iteration count, controlling the loop rhythm.

[0044] The invention will be further described below from several aspects, including system architecture, model establishment, design method, and effectiveness verification.

[0045] (1) Upper-level distribution network optimization model 1. Objective function To balance the economic efficiency and low-carbon operation of the distribution network, the upper-level model uses the overall system operating cost. The target is the lowest, including the cost of curtailment. Electricity purchase cost from the upper-level power grid Energy storage costs Gas turbine cost Carbon emission costs The specific formula is shown below. 1) Costs of wind and solar power curtailment in: and These represent the output power of the wind turbine and the photovoltaic system at time t, respectively. , These represent the maximum output power of the wind turbine and the photovoltaic system at time t, respectively. , These are the penalty cost coefficients for wind and solar power curtailment, respectively. , These are collections of wind turbines and photovoltaic units, respectively. 2) Electricity purchase cost from the upstream power grid , in: Let be the time-of-use electricity price at time t; Let t be the power purchased by the distribution network from the upper-level grid.

[0046] 3) Energy storage costs , in: This is the operating cost coefficient for the energy storage system. Let t be the charging power of the energy storage system at time t; Let t be the discharge power of the energy storage system at time t.

[0047] 4) Gas turbine cost, in: This refers to the set of distribution network nodes connected to the gas turbine. , , These are the quadratic, linear, and constant coefficients of the power generation cost of the gas turbine at node i, respectively. Let t be the active power output of the gas turbine connected to node i at time t.

[0048] 5) Carbon emission costs , in: Carbon emission coefficient; , These represent the carbon emission intensity of the upstream power grid's electricity purchase and the energy storage system at time t, respectively. Let be the carbon emission intensity of the gas turbine connected to node i.

[0049] The expression for the lowest overall operating cost of the upper-level system is: .

[0050] 2. Constraints To ensure the safe and stable operation of the distribution network, the model must meet the constraints of main grid power purchase, photovoltaic and wind turbine operation, gas turbine operation, energy storage, and power flow.

[0051] 1) Constraints on main grid power purchase in, and They are respectively t The maximum and minimum power purchases by the distribution network from the upstream grid at any given time; The constraints of the photovoltaic wind turbine. In the formula, and The first i node t Solar and wind turbine output at all times; and These represent the maximum output values ​​of the photovoltaic and wind turbines at time t, respectively. The specific expression for the gas turbine constraint is as follows: In the formula, and They are respectively connected to the first i The maximum and minimum active power output of the gas turbine at each node; A 0-1 variable representing the start-up and shutdown status of the unit; and The first The maximum ramp rate and maximum landslide rate of each node unit when it is in the start-up state.

[0052] The specific expression for the energy storage constraint is as follows: Energy storage operation should meet SoC constraints, including energy conversion constraints, capacity constraints, and daily power balance constraints.

[0053] In the formula: This represents the maximum charging and discharging power of the energy storage system. and These are 0 / 1 variables representing the charging / discharging state of energy storage; This refers to the rated capacity of the energy storage device; and These are the minimum and maximum capacities of the energy storage device, respectively. and These are the minimum and maximum values ​​of the state of charge, respectively; and These are the charging and discharging efficiencies of energy storage, respectively. and These represent the storage capacity at the beginning and end of a day, respectively.

[0054] The power flow constraint of the distribution network adopts the Distflow branch power flow model of the distribution network, and uses the second-order cone relaxation technique to convexify the nonlinear terms of voltage and current. The specific expression is as follows: In the formula: For i The set of the first nodes of the tail node; For i The set of tail nodes that is the first node; , respectively t time i Active and reactive power injected into the node; , They are respectively t From the node i Flow to Node j Active and reactive power; , They are respectively t From the node j Flow to Node k Active and reactive power; , Branch roads ij Resistance and reactance; and They are time t respectively i Node voltage and ij The square of the amplitude of the branch current; Node voltage and branch current constraints: In the formula: , They are respectively i Minimum and maximum values ​​of node voltage amplitude; , They are respectively via branch roads ij Minimum and maximum values ​​of current amplitude; The node power balance constraint: In the formula: , , , , and They are respectively t time i The active load of the node, the active power output of photovoltaic power, the active power output of wind power, the active power of energy storage, the active power obtained from the main grid, and the output of gas turbine.

[0055] (2) Fast calculation model for nodal marginal carbon potential The fast calculation model of node marginal carbon potential is based on the economic dispatch model of distribution network with distributed power sources. It uses the Lagrange multiplier method, dual vector and complementary relaxation conditions to establish an equivalent model. It uses the parameter differential perturbation method to derive and establish an analytical expression to accurately quantify the marginal relationship between the unit increment of node load and the overall carbon emission increment of the system. The compact form of the fast calculation model for the nodal marginal carbon potential is expressed as follows: Where A and b are the quadratic coefficient matrix and linear coefficient vector of the power generation cost, respectively, and the coefficient corresponding to the new energy power station is 0; It is the power output vector formed by the power source. E, D, and h are the corresponding coefficient matrices or vectors, and , ; Let be the dual vector of the constraint. This represents the active power output vector of all generating units in the distribution network.

[0056] Based on the Lagrange multiplier method and the Karush-Kuhn-Tucker condition, the Lagrange function is: The Lagrange equation is: The complementary relaxation condition is: in," "Indicates Hadamardi; When the load at a node changes slightly, the rate of change in the overall carbon emissions of the system is defined as the marginal carbon potential of the node, which can be expressed by the formula: in, This represents the total carbon emissions of the power distribution network. This represents the active power output vector of all generating units in the distribution network. It is by , and The power output vector formed by the power source ; This is the active power output vector of the higher-level system. This represents the active power output vector of all new energy power stations across all time periods. Let be the carbon emission factor of the i-th unit. Let i be the active power output vector of the i-th generating unit in the distribution network. The marginal carbon emission factor of the upper-level power grid; For the above This can be deduced as: in ; For the aforementioned incremental and The parameter differential perturbation method can be used to... The solution is performed, and the specific derivation is as follows: Differentiating the Lagrange equation yields: in Therefore, we can obtain From the above The derivation yields: We can obtain: Finally, we can seek the aforementioned... for: In conclusion: (3) Lower-level user model 1. Objective function The overall benefits on the user side consist of user energy efficiency, energy purchase costs, and low-carbon response incentives. The objective function can be expressed as: 1) User energy efficiency function To quantify the satisfaction that electricity consumption brings to users, a quadratic function is used to describe the relationship between energy efficiency and actual load.

[25] : In the formula: for t Immediate demand response afterload; and The electricity efficiency coefficient for users; 2) Energy purchase cost function Energy purchase cost refers to the fee paid by users to power distribution network operators for purchasing electricity, and it is affected by the time-of-use pricing strategy set by the higher authorities. In the formula: Developed for the upper-level system t The transaction price at any given moment; 3) Low-carbon response excitation function Users actively adjust their energy consumption behavior based on the received real-time DLMCI signals, thereby obtaining low-carbon incentive benefits: In the formula: for t Time Node i The load following demand response; 2. Constraints 1) Interruptible load Interruptible load constraints mainly include interruptible capacity, number of interruptions, maximum allowable interruption duration, and minimum time interval between two interruptions.

[0057] In the formula: for t Time Node i Reduced active power load; The maximum interruption amount for interruptible loads; To indicate t Time Node i A 0-1 variable indicating whether the load is reduced; For nodes i The maximum number of interruptible load interruptions; The maximum interruption duration for interruptible loads; The minimum interruption time interval for interruptible loads.

[0058] 2) Transferable load In the formula: and They are respectively t Time Node i Active load transferred in and out; and These are the maximum amount of load that can be transferred in and the maximum amount of load that can be transferred out, respectively. for t Time Node i 0-1 variables representing transferable loads; 3) Demand response power balance The demand response power balance constraint consists of interruptible load constraints and transferable load constraints. The demand response power balance constraint is as follows: .

[0059] Example 3: like Figure 3 The diagram shown is an IEEE 3-node topology. The gas turbine carbon potential is 0.48 kg CO2 / kWh. The ESS has a charge / discharge efficiency of 0.95, a state of charge upper and lower limits of 0.9 and 0.2, and initial carbon potentials of 0.3 kg CO2 / kWh and 0.4 kg CO2 / kWh, respectively. Nodes 8, 12, 24, and 25 are connected to interruptible loads, with an interruptibility limit of 20% of the current load at each node. Nodes 6, 21, 29, and 31 are connected to transferable loads, with a transferability limit of 10% of the current load at each node.

[0060] Example 4: To verify the effectiveness of the proposed low-carbon demand response mechanism based on nodal marginal carbon potential, this section sets up three comparative scenarios for analysis.

[0061] Scenario 1: Basic demand response scenario considering traditional time-of-use pricing mechanisms. This scenario does not consider carbon emission guidance mechanisms, but only utilizes the price differences between different time periods to drive flexible load shifting and reduction responses on the user side.

[0062] Scenario 2: Considering a low-carbon demand response scenario guided by the average carbon potential of nodes. This scenario utilizes time-sharing node carbon potential to guide users to achieve low-carbon energy use.

[0063] Scenario 3: Low-carbon demand response scenario considering the nodal marginal carbon potential mechanism. This scenario utilizes dynamic DLMCI signals to guide users' low-carbon energy consumption behavior.

[0064] like Figure 4 The figure shows the wind and solar power output prediction curves, load curves, and carbon potential of the upstream grid nodes for the entire system.

[0065] like Figure 5 The figure shows the DLMCI heat map of each node of the distribution network for 24 hours, calculated based on the output and load curves of wind power and photovoltaic power.

[0066] like Figure 7 The results show the demand response to load reduction: interruptible loads connected to nodes 8, 12, 24, and 25 performed significant reductions during the two main peak periods of the system, 8:00-11:00 and 18:00-21:00.

[0067] like Figure 8 The results show the demand response of transferable loads: the transferable loads connected to nodes 6, 21, 29 and 31 exhibit typical "peak avoidance and valley utilization" and "wind and solar load" characteristics.

[0068] like Figure 6 The diagram shown is the power balance diagram for scenario 3. Figure 9The power balance diagram for Scenario 1: Under the master-slave game optimization of price and environmental signals, Scenario 3 effectively guides the coordinated interaction between flexible loads, energy storage, and gas turbines, reduces the high-priced electricity purchases from the main grid by the distribution network during peak hours, and reasonably avoids long-distance power flow transmission, so that the network loss cost, energy storage operation cost, and overall cost of the system are all kept at a low level.

[0069] like Figure 10 The diagram shown is the power balance diagram for scenario 2. The comparison shows that the distribution of demand response actions, the transfer and shedding of loads, and the time evolution characteristics of the system's carbon potential are precisely matched in scenario 3.

[0070] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A method for low-carbon optimization scheduling of an active distribution network that takes into account the marginal carbon potential demand response of nodes, characterized in that, Includes the following steps: Step S1: Establish a two-layer collaborative low-carbon dispatching framework for the distribution network and users, including an upper-layer distribution network optimization dispatching model and a lower-layer low-carbon demand response model considering DLMCI; and establish a fast calculation model for node marginal carbon potential. Step S2: Calculate the lowest overall operating cost of the upper-level system using the upper-level distribution network optimization scheduling model; Step S3: Accurately quantify the DLMCI using the fast calculation model of the node marginal carbon potential; the DLMCI is the marginal relationship between the unit increment of node load and the overall carbon emission increment of the system. Step S4: Calculate the maximum comprehensive user-side benefits using the lower-level low-carbon demand response model that considers DLMCI. Step S5: Establish a two-layer optimization model and obtain an operable range containing the optimal operating state of the target through the bisection mechanism; Step S6: Repeat steps S2 to S5 until the length of the operable interval is less than the preset convergence accuracy, determine that the model has converged, and output the final scheduling result.

2. The active distribution network low-carbon optimization scheduling method considering nodal marginal carbon potential demand response as described in claim 1, characterized in that, The lowest overall operating cost of the upper-level system includes the cost of power curtailment. Electricity purchase cost from the upper-level power grid Energy storage costs Gas turbine cost and carbon emission costs .

3. The active distribution network low-carbon optimization scheduling method considering nodal marginal carbon potential demand response as described in claim 2, characterized in that, The cost of wind and solar power curtailment The specific calculations are as follows: in: and These represent the output power of the wind turbine and the photovoltaic system at time t, respectively. , These represent the maximum output power of the wind turbine and the photovoltaic system at time t, respectively. , These are the penalty cost coefficients for wind and solar power curtailment, respectively. , These are collections of wind turbines and photovoltaic units, respectively. The cost of electricity purchase from the upper-level power grid The specific calculations are as follows: in: Let be the time-of-use electricity price at time t; Let t be the power purchased by the distribution network from the upper-level grid; Energy storage costs The specific calculations are as follows: in: This is the operating cost coefficient for the energy storage system. Let t be the charging power of the energy storage system at time t; Let be the discharge power of the energy storage system at time t; The cost of the gas turbine The specific calculations are as follows: in: This refers to the set of distribution network nodes connected to the gas turbine. , , These are the quadratic, linear, and constant coefficients of the power generation cost of the gas turbine at node i, respectively. Let be the active power output of the gas turbine connected to node i at time t; The carbon emission cost The specific calculations are as follows: in: Carbon emission coefficient; , These represent the carbon emission intensity of the upstream power grid's electricity purchase and the energy storage system at time t, respectively. The carbon emission intensity of the gas turbine connected to node i; The minimum overall operating cost of the upper-level system is calculated as follows: 。 4. The active distribution network low-carbon optimization scheduling method considering nodal marginal carbon potential demand response as described in claim 1, characterized in that, The upper-level distribution network optimization scheduling model also includes main grid power purchase constraints, distributed unit operation constraints, energy storage charging and discharging constraints, distribution network power flow constraints, and node power balance constraints.

5. The active distribution network low-carbon optimization scheduling method considering nodal marginal carbon potential demand response as described in claim 4, characterized in that, The fast calculation model for the marginal carbon potential of nodes is based on the economic dispatch model of distribution networks with distributed power sources. It uses the Lagrange multiplier method, dual vectors and complementary relaxation conditions to establish an equivalent model. It uses the parameter differential perturbation method to derive and establish an analytical expression to accurately quantify the marginal relationship between the unit increment of node load and the overall carbon emission increment of the system. The compact form of the fast calculation model for the nodal marginal carbon potential is expressed as follows: Where A and b are the quadratic coefficient matrix and linear coefficient vector of the power generation cost, respectively, and the coefficient corresponding to the new energy power station is 0; It is the power output vector formed by the power source. E, D, and h are the corresponding coefficient matrices or vectors, and , ; Let be the dual vector of the constraint. This represents the active power output vector of all generating units in the distribution network.

6. The active distribution network low-carbon optimization scheduling method considering nodal marginal carbon potential demand response as described in claim 5, characterized in that, Based on the Lagrange multiplier method and the Karush-Kuhn-Tucker condition, the compact form of the Lagrange function representing the lowest overall operating cost of the upper-level system is: The Lagrange equation is: The complementary relaxation condition is: in," "Indicates Hadama accumulation; The rate of change in the overall carbon emissions of the system when a node experiences a small change in load is defined as the node marginal carbon potential, expressed by the formula: in, This represents the total carbon emissions of the power distribution network. This represents the active power output vector of all generating units in the distribution network. It is by , and The power output vector formed by the power source ; This is the active power output vector of the higher-level system. This represents the active power output vector of all new energy power stations across all time periods. Let be the carbon emission factor of the i-th unit. Let i be the active power output vector of the i-th generating unit in the distribution network. The marginal carbon emission factor of the upper-level power grid; For the above The derivation is as follows: in ; For the aforementioned incremental and Using the parameter differential perturbation method to The solution is performed, and the specific derivation is as follows: Differentiating the Lagrange equation yields: in Therefore, we can obtain From the above The derivation yields: We can obtain: Finally, we can seek the aforementioned... for: In conclusion: 。 7. The active distribution network low-carbon optimization scheduling method considering nodal marginal carbon potential demand response as described in claim 1, characterized in that, The greatest overall benefits on the user side include: user energy efficiency, energy purchase costs, and low-carbon response incentives.

8. The active distribution network low-carbon optimization scheduling method considering nodal marginal carbon potential demand response as described in claim 7, characterized in that, The user's energy efficiency can be expressed as: : in: for t Immediate demand response afterload; and The electricity efficiency coefficient for users; Energy purchase cost It can be represented as: in: Developed for the upper-level system t The transaction price at any given moment; The low-carbon response incentive It can be represented as in: for t Time Node i Load after demand response The objective function for maximizing overall user-side benefits is expressed as follows: 。 9. The active distribution network low-carbon optimization scheduling method considering nodal marginal carbon potential demand response as described in claim 1, characterized in that, The maximum user-side comprehensive benefit is subject to constraints, including interruptible load constraints, transferable load constraints, and demand response power balance constraints. The interruptible load constraints include interruptible capacity, number of interruptions, maximum allowable interruption duration, and minimum time interval between two interruptions; in: for t Time Node i Reduced active power load; The maximum interruption amount for interruptible loads; To indicate t Time Node i A 0-1 variable indicating whether the load is reduced; For nodes i The maximum number of interruptible load interruptions; The maximum interruption duration for interruptible loads; The minimum interruption interval for interruptible loads; The transferable load comfort includes, in: and They are respectively t Time Node i Active load transferred in and out; and These are the maximum amount of load that can be transferred in and the maximum amount of load that can be transferred out, respectively. for t Time Node i 0-1 variables representing transferable loads; The demand response power balance constraint consists of interruptible load constraints and transferable load constraints. The demand response power balance constraint is as follows: 。 10. The active distribution network low-carbon optimization scheduling method considering nodal marginal carbon potential demand response as described in claim 1, characterized in that, Steps S5 and S6 specifically include: Initialize the operable range of the target's optimal operating state; After each double-layer iteration, based on the comparison between the upper-layer objective function value and the midpoint objective value of the interval, if the current value is better than the midpoint value, the corresponding sub-interval is retained; otherwise, it is discarded. By continuously narrowing the search interval, the oscillations of the DLMCI signal and load response are forcibly suppressed until the interval length is less than the preset convergence accuracy, at which point the model is determined to have converged. After convergence, the final scheduling results are output: the upper-level main grid power purchase plan, energy storage charging and discharging strategy, and gas turbine output curve; the lower-level interruptible and transferable load response curves; and the carbon emissions and comprehensive operating cost indicators of the entire system.

Images