Multi-agent energy-carbon collaborative scheduling method with park thermal energy grade stratified matching and full privacy protection

Abstract

The application discloses a park heat energy grade layering matching and full privacy protection multi-agent energy-carbon collaborative scheduling method, relates to the technical field of comprehensive energy system operation and control, and comprises the following steps: S1: according to the temperature range, heat energy is divided into multiple grade levels, and actual heat of different grades is converted into equivalent work for representing the quality value; S2: a multi-agent energy-carbon collaborative optimization scheduling model is constructed, which comprises a local optimization objective function of each agent, a local constraint condition and a multi-agent global coupling constraint; S3: an alternating direction multiplier method is adopted, and only coupling variables are exchanged between each agent and a park collaborative layer for iterative solving, so that the multi-agent energy-carbon collaborative optimization scheduling model is solved under the premise of protecting the internal data privacy of each agent, and an optimal collaborative scheduling strategy is obtained. Therefore, the overall heat energy utilization efficiency of the park is significantly improved, and the collaborative optimization of energy efficiency and carbon emission reduction is realized.

Description

Technical Field

[0001] This invention belongs to the field of integrated energy system operation and control technology, specifically involving a multi-entity energy and carbon collaborative scheduling method for hierarchical matching of thermal energy quality and full privacy protection in industrial parks. More specifically, it is an optimized scheduling method for industrial parks that integrates cascade utilization of thermal energy and multi-entity energy and carbon collaboration. Background Technology

[0002] With the in-depth advancement of the construction of new power systems, industrial parks, as the core carriers of end-use energy consumption and carbon emissions, have become a key form for achieving high-proportion new energy consumption, coordinated energy and carbon management, and safe and efficient energy supply through their integrated energy systems of source, grid, load and storage. The industry's operational requirements have been upgraded from a single balance of power supply and demand to a multi-dimensional optimization goal of coordinated energy flow of electricity, heat and gas, cross-entity energy and carbon trading, and deep coupling of energy efficiency and carbon emission reduction.

[0003] The invention patent with publication number CN120013192A discloses an energy optimization scheduling method and system for industrial parks with electricity and carbon synergy. This scheme establishes a demand response model for electricity consumption and transferable load in the entire production process, constructs a price system that couples time-of-use electricity prices and floating carbon prices, adopts a two-layer centralized architecture of "park price decision-making - user energy optimization", and obtains the optimal electricity-carbon coupling price and power allocation through an improved particle swarm optimization algorithm, realizing electricity-carbon synergy scheduling and carbon value transmission under a single operating entity.

[0004] However, this solution and existing similar technologies still cannot meet all the actual needs of integrated energy systems in power industry industrial parks: they generally only focus on the synergistic optimization of electricity and carbon in the power dimension, completely ignoring the heat consumption of industrial parks, failing to distinguish the quality differences and value gradients of high-temperature process heat, medium-temperature production heat, and low-temperature heating heat, and failing to establish a graded cross-entity waste heat trading and precise matching mechanism; at the same time, they all rely on a centralized scheduling architecture, forcibly requiring each industrial entity to upload core sensitive information such as production energy consumption data, load characteristics, and cost functions, making it difficult to promote the technology on a large scale; in addition, heuristic algorithms have the defects of poor convergence and easy to get trapped in local optima, and only carry out static scheduling on a single time scale based on deterministic prediction data, without fully considering the typical uncertainties of power systems such as new energy output, load, and carbon price fluctuations, and cannot guarantee the safe and stable operation of the system under high proportion of new energy access.

[0005] Therefore, this invention aims to address the aforementioned industry pain points of inefficient and wasteful energy use and data privacy leaks, and provides a multi-entity energy and carbon collaborative scheduling method for industrial parks based on thermal energy grade hierarchical matching and adapting to the uncertainty of high proportion of new energy sources. Summary of the Invention

[0006] To overcome the shortcomings of existing technologies, this invention aims to provide a multi-entity energy and carbon collaborative scheduling method for industrial parks based on thermal energy grade matching. This method involves constructing a thermal energy grade classification and equivalent power conversion model, establishing a cross-entity waste heat matching mechanism, incorporating energy efficiency improvement and carbon trading into a unified optimization framework, and using distributed optimization technology to protect the data privacy of each entity, thereby achieving collaborative optimization of energy utilization efficiency and carbon emission control in industrial parks.

[0007] In a first aspect, embodiments of the present invention provide a multi-entity energy and carbon collaborative scheduling method for stratified matching of thermal energy quality in industrial parks and full privacy protection, comprising the following steps:

[0008] S1: Based on the temperature range, thermal energy is divided into multiple grade levels. For each grade level, the energy quality coefficient based on the second law of thermodynamics is used to convert the actual heat of different grades into equivalent work to characterize their quality value. The energy quality coefficient is determined based on the ratio of the heat source temperature to the ambient temperature.

[0009] S2: Construct a multi-agent energy and carbon collaborative optimization scheduling model that includes local optimization objective functions, local constraints, and multi-agent global coupling constraints for each agent. In the local optimization objective function, energy efficiency evaluation and cost accounting are characterized by the equivalent work output. The local constraints include energy balance constraints based on actual heat output and heat grade matching constraints based on temperature range and supply-demand temperature difference. The multi-agent global coupling constraints include at least the graded heat trading volume among the agents.

[0010] S3: The alternating direction multiplier method is adopted, and the coupling variables between each subject and the park collaboration layer are exchanged for iterative solution. Under the premise of protecting the internal data privacy of each subject, the multi-subject energy and carbon collaborative optimization scheduling model is solved to obtain the optimal collaborative scheduling strategy; wherein, the coupling variables include at least the cross-subject heat trading volume of the grade.

[0011] Furthermore, step S1 includes:

[0012] The heat energy grade classification divides heat energy into high temperature grade, medium temperature grade and low temperature grade according to the industrial heat temperature range;

[0013] The energy-mass coefficient, based on the second law of thermodynamics, is calculated using the following formula to convert the actual heat of different grades into equivalent work to characterize their quality value:

[0014] , ;in, Energy mass coefficient, Where T is the ambient temperature, H is the heat source temperature, and H is the initial heat. Equivalent work;

[0015] Furthermore, in the optimization model, the energy balance constraint uses the original heat.

[0016] Furthermore, step S1 also includes: establishing an engineering constraint model for cross-subject waste heat transport, wherein the engineering constraint model includes a transport distance constraint for limiting the transport distance, a temperature drop model for calculating temperature loss during transport, a heat exchange efficiency constraint for characterizing the heat exchanger efficiency, and a heat flow-temperature-pressure coupling constraint for relating heat, temperature and flow rate.

[0017] Furthermore, in step S2, the local optimization objective function of each subject aims to minimize the comprehensive cost of the subject within a single scheduling cycle. The comprehensive cost includes energy procurement cost, which represents the cost of purchasing electricity and gas; operation and maintenance cost, which represents the cost of operating and maintaining equipment; carbon emission cost, which represents the cost of carbon quota trading; and green transaction cost, which represents the cost of trading green electricity and green certificates.

[0018] Furthermore, in step S2, the thermal energy grade matching constraint includes:

[0019] The actual supply temperature of the heat flow must meet the temperature range corresponding to its grade;

[0020] The actual supply temperature of the heat flow must be greater than or equal to the sum of the required temperature of the heat load at the receiving end and the preset minimum heat exchange temperature difference;

[0021] Heat-temperature-flow coupling equations are used to correlate heat, temperature, and flow rate.

[0022] Furthermore, the heat-temperature-flow coupling equation is a nonlinear constraint, which is transformed into a linear constraint through a piecewise linearization method. The transformed model is a multi-agent energy and carbon collaborative optimization scheduling model in the form of mixed integer linear programming (MILP).

[0023] Furthermore, step S3 specifically includes:

[0024] Initialize the number of iterations, the initial values ​​of the coupling variables of each subject, and the Lagrange multipliers;

[0025] After receiving the coupling variables, each entity independently solves its own local optimization subproblem to update its local runtime plan and coupling variables;

[0026] Each entity will send the updated coupling variables to the park's collaboration layer;

[0027] The park collaboration layer calculates the residual of the global coupling constraint based on the coupling variables sent by each subject. If the residual is less than the preset convergence threshold, the iteration is terminated. Otherwise, the Lagrange multipliers are updated based on the residual and the iteration returns to the step of independently solving their respective local optimization subproblems.

[0028] Furthermore, step S3 also includes a convergence acceleration strategy, which includes at least one of the following strategies: an adaptive penalty factor adjustment strategy, which dynamically adjusts the penalty factor according to the ratio of the original residual to the dual residual; an early stopping mechanism, which terminates the iteration in advance when the cost reduction is less than a preset value; and a hot start strategy, which uses the optimization result of the previous scheduling cycle as the initial solution of the current cycle.

[0029] Furthermore, the method also includes step S4: using a stochastic optimization method to handle the uncertainties of new energy output, load, and market prices, specifically including:

[0030] A multi-stage stochastic programming framework is adopted. Multiple uncertainty scenarios are obtained by sampling and clustering reduction. The scheduling problem is decomposed into day-ahead and day-intraday stages, and step S3 is called to solve the problem under each uncertainty scenario.

[0031] Furthermore, the multi-stage stochastic programming framework adopts stochastic dual dynamic programming as the outer framework and step S3 as the inner solver. The stochastic dual dynamic programming is used to perform cross-scenario coordination optimization and update state variables. Step S3 is used to solve the multi-agent scheduling sub-problem under each deterministic scenario. The two are nested and iterated until convergence.

[0032] Secondly, embodiments of the present invention also provide an electronic device, the electronic device comprising:

[0033] One or more processors;

[0034] Storage device for storing one or more programs;

[0035] When the one or more programs are executed by the one or more processors, the one or more processors implement the multi-entity energy and carbon collaborative scheduling method with stratified matching of park thermal energy grade and full privacy protection as described in any embodiment of the present invention.

[0036] Thirdly, embodiments of the present invention also provide a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the multi-entity energy and carbon collaborative scheduling method for stratified matching of park thermal energy grade and full privacy protection as described in any embodiment of the present invention.

[0037] Compared with existing technologies, the present invention achieves the following beneficial effects:

[0038] (1) This invention can significantly improve the overall thermal energy utilization efficiency of the park by matching thermal energy grade and utilizing waste heat across entities. Compared with the traditional centralized scheduling scheme that does not distinguish thermal energy grade, simulation results show that thermal energy utilization efficiency can be greatly improved, thereby reducing the consumption of purchased energy.

[0039] (2) This invention incorporates energy efficiency improvement, carbon emission control, and green electricity consumption into a unified optimization framework. Through the cascade utilization of thermal energy and the collaboration of multiple entities, it helps to reduce the overall carbon emission level of the park and achieve synergistic optimization of energy efficiency and carbon emission reduction.

[0040] (3) The present invention adopts the distributed optimization architecture of the alternating direction multiplier method (ADMM). The core data such as equipment parameters, cost functions, and load curves of each subject are kept locally throughout the process. Only necessary coupling variables (such as the heat trading volume of different grades and the green electricity trading volume) need to be exchanged. This effectively solves the data privacy protection problem in multi-subject collaboration and has good engineering promotion value.

[0041] (4) The mixed-integer linear programming (MILP) model constructed in this invention, combined with a distributed solution strategy, can achieve acceptable solution efficiency under typical park scale, meeting the actual time requirements of day-ahead scheduling. At the same time, the introduced convergence acceleration strategies such as adaptive penalty factor adjustment and early stopping mechanism further improve the stability and practicality of the algorithm in engineering applications. Attached Figure Description

[0042] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings. The drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0043] Figure 1 This is a flowchart of the multi-entity energy and carbon collaborative scheduling method with hierarchical matching of park thermal energy grade and full privacy protection provided in the embodiments of the present invention;

[0044] Figure 2 This is a schematic diagram of thermal energy grade classification and equivalent work conversion according to an embodiment of the present invention;

[0045] Figure 3 This is a flowchart of the distributed collaborative solution iterative process based on ADMM according to an embodiment of the present invention;

[0046] Figure 4 This is a diagram of the nested solution architecture of multi-stage stochastic programming and ADMM according to an embodiment of the present invention;

[0047] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0048] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, the accompanying drawings show only the parts relevant to the present invention, and not all of the structures.

[0049] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe operations (or steps) as being processed sequentially, many of these operations (or steps) may be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations may be rearranged. The process may be terminated when its operation is completed, but may also have additional steps not included in the figures. The process may correspond to a method, function, procedure, subroutine, subroutine, etc.

[0050] Example 1

[0051] like Figure 1 The diagram shows a flowchart of a multi-entity energy and carbon collaborative scheduling method with hierarchical matching of park thermal energy quality and full privacy protection provided in Embodiment 1 of the present invention. The method specifically includes the following steps:

[0052] S1: Based on the temperature range, thermal energy is divided into multiple grade levels. For each grade level, the energy quality coefficient based on the second law of thermodynamics is used to convert the actual heat of different grades into equivalent work to characterize their quality value. The energy quality coefficient is determined based on the ratio of the heat source temperature to the ambient temperature.

[0053] Step S1 is used to construct a multi-entity thermal energy grade classification and equivalent conversion model for the industrial park. This step provides a quantitative basis for thermal energy matching, including: the thermal energy grade classification divides thermal energy into high-temperature, medium-temperature, and low-temperature levels according to the industrial heating temperature range. For example... Figure 2 The diagram shown is a schematic representation of the thermal energy grade classification and equivalent work conversion according to an embodiment of the present invention. The specific steps are as follows:

[0054] S1.1, Definition of heat energy grade classification and heat flow variable:

[0055] Based on the common temperature ranges for heat use in industry, thermal energy is divided into three levels: high-temperature heat (≥150℃), medium-temperature heat (60℃-150℃), and low-temperature heat (<60℃). This classification standard refers to the relevant provisions of the "Guidelines for Comprehensive Utilization of Industrial Waste Heat in Stages" (GB / T39091-2020).

[0056] To support multi-product, multi-chain scenarios, the following variables are defined:

[0057] :main body To the main body During the period The actual heat delivered in Tier 1 (kWh), Tier These correspond to high, medium, and low temperatures, respectively, and serve as variables for optimizing decision-making.

[0058] The actual supply temperature (°C) of the corresponding heat flow is used as an auxiliary decision variable.

[0059] : Volumetric flow rate corresponding to heat flux ( ), which are auxiliary decision variables.

[0060] Explanation of variable relationships: Tier only represents the heat quality category; the actual temperature is determined by the variable. This indicates that the corresponding grade range constraint is met (e.g., when Tier=H). This approach retains the convenience of gradation classification while avoiding redundant conflicts between categorical variables and continuous temperature variables.

[0061] S1.2 Equivalent work conversion based on energy mass coefficient:

[0062] By introducing the energy-mass coefficient based on the second law of thermodynamics, an equivalent work conversion model for thermal energy of different grades is established. Energy-mass coefficient The calculation formula is:

[0063]

[0064] in, Let T be the ambient temperature (taken as 298 K), and T be the heat source temperature (K). It should be noted that the temperatures T and T in the formula... All values ​​are thermodynamic temperatures (unit: K). In practical engineering, if the measured temperature is in Celsius (t) (unit: °C), it needs to be calculated using the formula... Perform the conversion. Different grades of heat are converted into equivalent work using the following formula:

[0065]

[0066] in, Original heat (kWh) Equivalent work (kWh).

[0067] Thermodynamic consistency note: In the optimization model, the energy balance constraint uses actual heat. (Satisfying the first law of thermodynamics, i.e., conservation of energy), the energy efficiency evaluation and carbon accounting in the objective function use equivalent work. (This reflects the quality difference in the second law of thermodynamics). The two are related through the above conversion formula, a practice consistent with the principles of (exergy) analysis. The portion of equipment operation and maintenance costs related to thermal output is calculated based on equivalent work. The cost is adjusted to reflect the higher equipment wear and maintenance costs associated with generating high-grade heat energy; other fixed costs (such as equipment depreciation) are still amortized based on actual operating time.

[0068] S1.3 Constraints on cross-body waste heat transfer engineering:

[0069] Step S1.3 is used to establish an engineering constraint model for cross-body waste heat transport. This model includes at least one of the following: a transport distance constraint to limit the transport distance; a temperature drop model to calculate temperature loss during transport; a heat exchange efficiency constraint to characterize heat exchanger efficiency; and a heat flow-temperature-pressure coupling constraint to correlate heat, temperature, and flow rate. Specifically, establishing the engineering constraint model for waste heat transport includes:

[0070] (1) Conveying distance constraints: The maximum economical conveying distance for different grades is set according to the type of heat medium and the insulation level of the pipeline network. High-temperature heat (steam) insulation costs are high but the loss per kilometer is small, and the economical distance can reach 5 km; medium and low temperature heat (hot water) is limited by pump power consumption and heat dissipation loss, and the economical distance is usually ≤2 km. In this embodiment, 2 km is uniformly taken as a conservative value;

[0071] (2) Temperature drop model: linear approximation is adopted ,in, Temperature loss (°C) during heat transfer; Temperature drop coefficient per unit distance (°C / km) reflects the thermal insulation performance and heat dissipation characteristics of the pipeline; : Pipeline length for heat flow transport (km). This linear model is based on offline simulation fitting of a typical industrial steam pipeline (DN200, insulation thickness 100 mm, flow velocity 25 m / s). Under a relatively stable flow scheduling timescale (1 hour), the error compared to the detailed thermodynamic model is less than 5%, which is acceptable for engineering applications. Typical industrial heating network The value ranges from 0.5 to 2℃ / km, and 0.8℃ / km is used in this embodiment. In practical applications, the value can be calibrated according to the specific pipeline network parameters. value;

[0072] (3) Heat exchange efficiency constraint: After the heat energy is transported, it is supplied to the receiving end through the heat exchanger. The heat exchange efficiency of the heat exchanger is required. Take the range of 0.85 to 0.95;

[0073] (4) Heat flow-temperature-pressure coupling constraint: The actual heat transferred by the heat flow is related to the temperature and flow rate as follows: ,in The actual heat transferred by the heat flow (kWh) needs to be converted to kWh unit after calculation; Specific heat capacity of the heat transfer medium, taken as This represents the heat capacity of water or steam. The density of the heat transfer medium is taken as... , representing the density of water; Volumetric flow rate of heat flux ( ); : Temperature difference (°C) between the inlet and outlet of heat flow during heat exchange or transport. Unit conversion notes: If in the formula... Take 4.18 , Pick , Pick , If we take ℃, then the unit of the calculation result is ℃. The conversion to kWh requires dividing by 3600. All heat calculations in this invention are uniformly converted to kWh units. This model simplifies the physical properties of the heat medium to constants, ignores the nonlinear effects of pressure loss, and focuses on energy balance and grade matching at the dispatch level.

[0074] S2: Construct a multi-agent energy and carbon collaborative optimization scheduling model that includes local optimization objective functions, local constraints, and multi-agent global coupling constraints for each agent. In the local optimization objective function, energy efficiency evaluation and cost accounting are characterized by the equivalent work output. The local constraints include energy balance constraints based on actual heat output and heat grade matching constraints based on temperature range and supply-demand temperature difference. The multi-agent global coupling constraints include at least grade-based cross-agent heat trading total balance constraints.

[0075] Step S2 is used to establish a multi-entity energy and carbon collaborative optimization scheduling model. This step S2 integrates thermal energy grade matching and energy and carbon control into the optimization model.

[0076] S2.1 Definition of model decision variables:

[0077] (1) Energy procurement variables:

[0078] The main body in the time period Purchased power (kW);

[0079] The main body in the time period Gas purchase volume ( );

[0080] (2) Thermal energy interaction variables:

[0081] :main body To the main body During the period The actual heat delivered in Tier 1 Positive value indicates from Towards Send out;

[0082] : The supply temperature of the corresponding heat flow Auxiliary decision-making variables;

[0083] Volumetric flow rate corresponding to heat flux Auxiliary decision-making variables;

[0084] (3) Transaction variables:

[0085] :main body Green electricity trading volume (kWh, buys are positive, sells are negative);

[0086] :main body The trading volume of green certificates (in units, buys are positive, sells are negative);

[0087] :main body Carbon credit trading volume (tons, buy is positive, sell is negative);

[0088] (4) Internal device variables (local to each entity):

[0089] Power generation capacity of combined heat and power equipment (kW);

[0090] Heat output of cogeneration equipment (kWh);

[0091] Heat output of gas-fired boiler (kWh);

[0092] , Energy storage charging / discharging power (kW);

[0093] : Energy storage state of charge.

[0094] S2.2 Construct the local optimization objective function for each subject:

[0095] In step S2.2, the local optimization objective function of each entity aims to minimize the entity's overall cost within a single scheduling cycle. This overall cost includes energy procurement costs (representing electricity and gas purchase fees), operation and maintenance costs (representing equipment operation and maintenance costs), carbon emission costs (representing carbon quota trading fees), and green trading costs (representing green electricity and green certificate trading fees). Specifically, the optimization objective is to minimize the entity's overall cost within a single scheduling cycle.

[0096]

[0097] in: :main body The comprehensive cost (RMB / cycle) within a scheduling cycle includes energy procurement, equipment operation and maintenance, carbon emissions, and green electricity / green certificate trading costs; :main body Energy procurement costs (RMB / cycle), including electricity and gas purchase costs; :main body The equipment operation and maintenance cost (RMB / cycle) includes the operation and maintenance costs of equipment such as power generation, heat generation, energy storage and pumping; :main body The carbon emission cost (yuan / cycle) is calculated by multiplying the difference between the actual emissions and the allowance by the carbon price. :main body The transaction cost of green electricity and green certificates (yuan / cycle) is positive when buying and negative when selling.

[0098] Energy procurement costs: ,in, The main body in the time period Purchased power (kW); Time period Time-of-use electricity price (RMB / kWh); The main body in the time period gas purchase volume ; Natural gas unit price (yuan) );

[0099] Operation and maintenance costs: The part related to thermal output ( , According to equivalent work Conversion, coefficient Differentiate between different equipment types; pump power consumption is calculated based on flow rate and delivery distance. Unit operation and maintenance cost coefficient of combined heat and power equipment (yuan / kWh); Cogeneration equipment during the time period Power generation capacity (kW); The unit operation and maintenance cost coefficient (yuan / kWh) of the heat output of combined heat and power equipment, based on equivalent power. Conversion; Cogeneration equipment during the time period Heat production (kWh); The unit operation and maintenance cost coefficient (yuan / kWh) of the heat output of a gas-fired boiler, based on equivalent power. Conversion; Gas boilers during certain periods Heat output (kWh); : Unit operation and maintenance cost coefficient of energy storage equipment charging / discharging power (yuan / kWh); Energy storage devices during time periods The charging power (kW); Energy storage devices during time periods The discharge power (kW); Unit flow rate • distance operation and maintenance cost coefficient of pumping equipment ( ); :main body To the main body During the period Volumetric flow rate of Tier 1 heat flux delivered ( ); : Length of the pipeline for heat transfer (km);

[0100] Carbon emission costs: ,in, :main body Actual carbon emissions (tonnes) within a scheduling cycle ), calculated according to the IPCC guidelines; :main body Initial carbon emission allowances (tonnes) within a scheduling cycle ); Carbon market trading price (yuan / ton) );

[0101] Green electricity / green certificate transaction costs: .in, :main body The green electricity trading volume (kWh) is positive for buying and negative for selling; Green electricity trading price (RMB / kWh); :main body The trading volume (in units) of green certificates is positive for buying and negative for selling; : Green certificate trading price (RMB / certificate).

[0102] S2.3, Set local constraints for each subject:

[0103] (1) Equipment operating constraints:

[0104] Upper and lower limits of output: ;in, : Lower limit of equipment output (kW); Power output (kW) of the equipment during a certain period of time; : The upper limit of equipment output (kW);

[0105] Climbing speed constraint: ;in, Equipment during the time period Power output (kW); Equipment during the time period Power output (kW); The maximum ramp rate of the equipment (kW / time period) is the maximum allowable change in output between adjacent time periods.

[0106] Energy storage constraints: , .in, Energy storage devices during time periods State of charge at the end (kWh or percentage); Energy storage devices during time periods State of charge at the end (kWh or percentage); : Charging efficiency of energy storage devices; Energy storage devices during time periods The charging power (kW); Energy storage devices during time periods The discharge power (kW); The discharge efficiency of energy storage devices; : Lower limit of the state of charge (kWh or percentage) of energy storage devices; : The upper limit of the state of charge (kWh or percentage) of energy storage devices.

[0107] (2) Energy balance constraints (based on actual heat) ):

[0108] Power balance: ;in, The main body in the time period Purchased power (kW); The main body in the time period Photovoltaic power generation (kW); Cogeneration equipment during the time period Power generation capacity (kW); Energy storage devices during time periods The discharge power (kW); The main body in the time period The electrical load power (kW); Energy storage devices during time periods The charging power (kW); The main body in the time period Power sold to external customers (kW);

[0109] Thermal equilibrium: ;in, Cogeneration equipment during the time period Heat production (kWh); Gas boilers during certain periods Heat output (kWh); The main body in the time period Internal process waste heat recovery (kWh); Time period Inside, other entities To the main body The sum of the actual heat transferred in Tier 1 (kWh), i.e., the heat flowing into the main body Calories; The main body in the time period Heat load demand (kWh); The main body in the time period Heat loss in heating network transmission (kWh), according to calculate, This is the heat loss coefficient, taken as 0.05; The main body in the time period The total heat supply is (kWh). All heat is actual heat (kWh), satisfying the law of conservation of energy.

[0110] (3) Thermal energy grade matching constraints:

[0111] Furthermore, in step S2, the heat energy grade matching constraints include: the actual supply temperature of the heat flow must meet the temperature range corresponding to its grade level; the actual supply temperature of the heat flow must be greater than or equal to the sum of the required temperature of the receiving end heat load and the preset minimum heat exchange temperature difference; and a heat-temperature-flow coupling equation used to correlate heat, temperature, and flow rate. The heat-temperature-flow coupling equation is a nonlinear constraint, which is transformed into a linear constraint using a piecewise linearization method. The transformed model is a multi-agent energy and carbon collaborative optimization scheduling model in the form of mixed integer linear programming (MILP). Specifically:

[0112] Feasible temperature range: The temperature range corresponding to the grade must be met; that is, if Tier = H, then... (Converted to thermodynamic temperature approximately 423.15 K); if Tier = M, then If Tier = L, then ;

[0113] Heat exchange temperature difference constraint: , ;in, :main body To the main body The delivered heat flow during the time period Actual supply temperature (°C); Tier: Heat energy grade category, with values ​​of H (high temperature), M (medium temperature), and L (low temperature); The receiving end subject during the time period The required temperature for heat load (°C); Minimum heat exchange temperature difference, taken as 10℃;

[0114] Temperature-flow-heat coupling equation: ,in :main body To the main body During the period The actual heat transferred in the Tier 1 stage (kWh); c: specific heat capacity of the heat transfer medium, taken as... ; : Density of the heat transfer medium, take ; :main body To the main body During the period Volumetric flow rate of Tier 1 heat flux delivered ; The return water temperature of the heating network is taken as 60℃; divide by 3600 to convert kJ / h to kWh.

[0115] Linearization: Temperature-flow-heat coupling equations involve variable multiplication This falls under nonlinear constraints. This invention employs a piecewise linearization method: discretizing the temperature variable into... A range, introducing binary variables The product term is represented as The transformed model is in the form of mixed integer linear programming (MILP) and can be solved using commercial solvers. : A binary variable indicating whether the temperature falls within the specified range. Within a discrete interval; : No. The representative temperature value (°C) corresponding to each discrete temperature interval.

[0116] Carbon emission constraints: , constraint is .in, :main body Actual carbon emissions (tonnes) within a scheduling cycle ); The main body in the time period gas purchase volume ; The carbon emission factor of natural gas, taken as... ; The main body in the time period Purchased power (kW); : Power grid during the period carbon emission factors ( ); :main body The green electricity trading volume (kWh) is positive for buy orders; The carbon emission factor of green electricity, taken as ; :main body Initial carbon emission allowances (tonnes) within a scheduling cycle ); :main body Carbon trading volume (tons) Buying is positive, selling is negative;

[0117] Green energy consumption constraints: ,in (Buy is positive). The main body in the time period Actual green electricity consumption (kWh); The main body in the time period Photovoltaic power generation (kW); :main body The green electricity trading volume (kWh) shows a positive buy signal; The main body in the time period The electrical load power (kW).

[0118] S2.4, Set global coupling constraints for multiple subjects:

[0119] Cross-entity heat trading total balance (by grade): ;in, :main body To the main body During the period The actual heat (kWh) delivered in the Tier 1 level; Time period Inside, all main bodies To the main body The sum of Tier 1 heat (kWh) transported, i.e., the heat flowing into the main body Calories; Time period Inside, all subjects To the main body The sum of Tier 1 heat transferred (kWh), i.e. from the main body The heat that flows out; This balance constraint must be met for all three heat energy grades: high, medium, and low.

[0120] Balance of total green electricity trading volume: ;in, :main body The green electricity trading volume (kWh) is positive for buying and negative for selling; The sum of green electricity transactions by all entities within the park is zero, indicating that green electricity is only traded within the park, and the total volume is balanced.

[0121] Balance of total trading volume for green certificates / carbon rights: ,in, :main body The trading volume of green certificates or carbon rights (in units or tons) is positive for buying and negative for selling; If the sum of the green certificate or carbon credit trading volumes of all entities within the park is zero, it means that the trading only circulates within the park and the total volume is balanced.

[0122] S2.5 Complete Optimization Model Mathematical Representation (MILP Form):

[0123] In summary, the optimization model constructed in this invention can be expressed in the form of standard mixed-integer linear programming (MILP):

[0124]

[0125] in, The total number of entities participating in the scheduling within the park; : The total number of time periods within the scheduling cycle; Time period Time-of-use electricity price (RMB / kWh); :main body During the period Purchased power (kW); Natural gas unit price (yuan / ); :main body During the period gas purchase volume ; Summing over all device types; : Unit operation and maintenance cost coefficient for equipment type dev (yuan / kWh or equivalent unit); :main body During the period The output of the device dev (kWh or equivalent unit); Carbon market trading price (yuan / ton) ); :main body During the period Actual carbon emissions (tons) ); :main body During the period carbon emission allowances (tonnes) ); Green electricity trading price (RMB / kWh); :main body During the period The green electricity trading volume (kWh) is positive for buying and negative for selling; Green certificate trading price (RMB / certificate); :main body During the period The trading volume of green certificates (in units) is positive for buying and negative for selling.

[0126] The constraints include:

[0127] ① Equipment output constraints (linear inequalities);

[0128] ② Energy balance constraints (linear equations);

[0129] ③ Thermal energy quality matching constraints (including piecewise linearized temperature-flow coupling equations);

[0130] ④ Global coupling constraints (linear equations);

[0131] ⑤ Variable nonnegativity and integer constraints (device start-up and shutdown are 0-1 variables).

[0132] This model can be solved using mature toolchains such as YALMIP / Gurobi.

[0133] S3: The alternating direction multiplier method is adopted, and the coupling variables between each entity and the park collaboration layer are exchanged for iterative solution. Under the premise of protecting the internal data privacy of each entity, the multi-entity energy and carbon collaborative optimization scheduling model is solved to obtain the optimal collaborative scheduling strategy. Among them, the coupling variables include at least the graded heat trading volume between each entity.

[0134] Step S3 provides a collaborative solution method based on distributed optimization. Step S3 achieves model solving while protecting the data privacy of each entity.

[0135] S3.1 Decentralized solution architecture:

[0136] The Alternating Directional Multiplier Method (ADMM) is used for distributed solution. Each entity only needs to exchange coupling variables (cross-entity hot transaction volume by grade) with the park's collaborative layer. The core information of the green electricity / green certificate / carbon rights trading volume, including internal equipment parameters, cost functions, and load data, is retained locally throughout the entire process.

[0137] S3.2 Distributed Iterative Process:

[0138] Furthermore, such as Figure 3 The diagram shown is a flowchart of the distributed collaborative solution iterative process based on ADMM according to an embodiment of the present invention. Step S3 specifically includes:

[0139] Initialization: Set the number of iterations Initialize the initial values ​​of the coupling variables of each subject and set the convergence threshold. The initial values ​​of the Lagrange multipliers are set to 0, and the penalty factor is... Set the value to 1.0;

[0140] Local optimization: After receiving the coupling variables from neighboring entities, each entity independently solves the local optimization subproblem and updates its local execution plan and the coupling variables for external interactions;

[0141] Information exchange: Each entity sends the updated coupling variables to the park's collaboration layer;

[0142] Convergence judgment: The park collaboration layer calculates the residuals of global coupling constraints based on the coupling variables sent by each subject. If res < ε and the cost change over three consecutive iterations is less than ε, If the condition is met, the iteration terminates; otherwise, update the Lagrange multipliers and continue iterating.

[0143] The multiplier update formula adopts the standard ADMM form:

[0144]

[0145] in, : The current iteration number of the ADMM distributed iteration, initially set to 0; ε: The convergence threshold, set to... This is used to determine whether the global coupling constraint residuals satisfy the convergence condition. : The penalty factor of the ADMM algorithm, initially set to 1.0, used to balance the convergence speed of the original residual and the dual residual; res: The residual value of the global coupling constraint, which is the maximum value of the total deviation of each coupling variable; All subjects To the main body The sum of Tier 1 heat delivered, i.e., the total global heat volume at that grade. The absolute deviation of the total amount of heat traded across entities reflects the degree to which heat balance constraints are met. : The algebraic sum of the green electricity trading volumes of all entities; The absolute value of the deviation in the total volume of green electricity transactions reflects the degree to which the green electricity balance constraint is met. : The algebraic sum of the trading volumes of all green certificates or carbon rights by all entities; The absolute value of the deviation in the total trading volume of green certificates or carbon rights reflects the degree to which the trading balance constraint is met. : No. Lagrange multipliers in round-recursion iteration; : No. Lagrange multipliers in round-recursion iteration; : Penalty factor, which controls the step size of multiplier updates; : No. In each iteration, the subject's local decision variables (such as electricity purchased, heat generated, energy storage charging and discharging power, etc.). : No. In each iteration, the consistent target value of the coupled variables (such as the heat trading volume and green electricity trading volume agreed upon by each entity). : Lagrange multipliers, corresponding to the dual variables of global coupling constraints, used to penalize inconsistencies in coupling variables; Local decision variables include local optimization variables such as energy procurement, equipment output, energy storage charging and discharging, and heat trading volume; The consistency target value of the coupled variables serves as a consensus reference value for the local decision variables of each subject under global coupling constraints.

[0146] S3.3 Convergence Acceleration Strategy:

[0147] Furthermore, to improve engineering feasibility, a convergence acceleration strategy is introduced. This strategy includes at least one of the following: adaptive penalty factor adjustment (dynamically adjusting the penalty factor based on the ratio of the original residual to the dual residual), an early stopping mechanism (terminating the iteration early when the cost reduction is less than a preset value), and hot-starting (using the optimization result of the previous scheduling cycle as the initial solution for the current cycle). Specifically:

[0148] Adaptive penalty factor adjustment: dynamically adjusted based on the ratio of the original residual to the dual residual. To avoid shocks;

[0149] Early termination mechanism: If the cost decrease is less than 0.1% for 5 consecutive iterations, the process will be terminated early.

[0150] Warm start: Use the optimization result of the previous scheduling cycle as the initial solution for the current cycle.

[0151] Furthermore, the method in this embodiment of the invention also includes step S4: using a stochastic optimization method to handle the uncertainties of new energy output, load, and market prices. This step S4 is used to address fluctuations in new energy output, load, and market prices. Specifically, it includes: using a multi-stage stochastic programming framework, generating and clustering multiple uncertainty scenarios through sampling, decomposing the scheduling problem into day-ahead and intraday stages, and calling step S3 to solve in each uncertainty scenario. The multi-stage stochastic programming framework uses stochastic dual dynamic programming as the outer framework and step S3 as the inner solver. Cross-scenario coordination optimization and state variable updates are performed through the stochastic dual dynamic programming. Step S3 solves the multi-agent scheduling sub-problem under each deterministic scenario, with the two nested iteratively until convergence. Figure 4 The diagram shown illustrates the nested solution architecture of multi-stage stochastic programming and ADMM according to an embodiment of the present invention. Specifically, it includes the following steps:

[0152] S4.1 Uncertainty Scenario Generation and Reduction:

[0153] Initial scenario sets for photovoltaic power output, load demand, and carbon price are generated using Latin hypercube sampling. The uncertainty distribution assumptions are as follows:

[0154] Photovoltaic output prediction error: Assumed to be a normal distribution with a mean of 0 and a standard deviation of 15% of the predicted value;

[0155] Load forecasting error: Assumed to be a normal distribution with a mean of 0 and a standard deviation of 5% of the forecast value;

[0156] Carbon price fluctuations: assumed to be geometric Brownian motion, with an annual volatility of 20%.

[0157] Based on the above distribution, 100 initial scenarios were generated, and then reduced to 10 core scenarios using k-means clustering to ensure coverage of more than 95% of the uncertainty features. The selection of the number of clusters was based on the silhouette score, which evaluated the clustering effect at different numbers of clusters. The silhouette score reached its maximum value of 0.62 when the number of clusters was 10, indicating the optimal clustering effect; therefore, 10 core scenarios were selected.

[0158] S4.2 Multi-stage stochastic programming framework:

[0159] Stochastic Dual Dynamic Programming (SDDP) is used as the outer framework to decompose the scheduling problem into a "day-to-day" time sequence. The day-to-day phase determines the medium- to long-term trading strategy and equipment start-up / shutdown plan (time step 1 hour); the intraday phase (rolling every 15 minutes) revises the operation plan based on the latest forecasts. The number of SDDP iterations is typically 20-50 (based on empirical values ​​from Shapiro, 2011); this embodiment uses 30 iterations.

[0160] S4.3 Nested solution structure:

[0161] Under each SDDP scenario node, the ADMM distributed optimization algorithm from step S3 is invoked to solve the multi-agent scheduling subproblem. SDDP is responsible for cross-scenario coordinated optimization, generating trajectories of state variables (energy storage capacity, carbon quota surplus) through forward simulation and updating the approximation of the cost-to-go function through backward recursion. ADMM is responsible for the multi-agent collaborative solution under each deterministic scenario. The two are nested and iterated until convergence.

[0162] S4.4 Computational Complexity Analysis and Engineering Feasibility Assurance:

[0163] Theoretical complexity: SDDP typically requires 20 to 50 iterations, with each iteration requiring the solution of 10 ADMM subproblems for different scenarios. Each ADMM subproblem contains 3 to 10 main MILP problems.

[0164] Engineering implementation strategy: Parallel computing is used to solve ADMM problems in different scenarios; integer variables (such as equipment start-up and shutdown) are decided in the day-ahead phase, and only continuous variables are optimized in the intraday phase; the distributed computing capabilities of commercial solvers such as Gurobi are used to accelerate the main solution.

[0165] Hardware requirements: The test environment consisted of an Intel Core i7-12700 CPU (12 cores), 32 GB RAM, and a Gurobi 10.0 Academic license. For actual project deployment, a server configuration of equivalent or higher is recommended.

[0166] Convergence guarantee: Set a maximum number of iterations (SDDP: 50 times, ADMM: 100 times) to ensure that the algorithm terminates within a finite amount of time;

[0167] Recommendations for practical deployment: The assumptions made in this method, such as the linear temperature drop model and the simplification of physical property constants, require recalibration of the pipeline network parameters using CFD or thermal simulation tools during actual engineering deployment to ensure model accuracy.

[0168] The uncertainties addressed in this invention (photovoltaic output, load, carbon price) are characterized by multi-stage (day-ahead to intraday) and time-series correlation. Compared to traditional single-stage stochastic programming or robust optimization, the SDDP framework can better characterize the coupling relationship between state variables such as energy storage and thermal storage at different scheduling stages, avoiding inaccurate state variable decisions due to scenario reduction, thereby obtaining a better multi-stage decision-making strategy while ensuring computational efficiency.

[0169] Step S5: Simulation Verification Method for Scheduling Strategies

[0170] S5.1 Simulation Platform and Parameter Settings:

[0171] A simulation model was built on the MATLAB R2022b platform, using the YALMIP toolbox and the Gurobi 10.0 solver. The hardware environment consisted of an Intel Core i7-12700 CPU and 32 GB of RAM. The scheduling cycle was 24 hours, with a time step of 1 hour.

[0172] S5.2 Test system parameters:

[0173] The test system consists of three main industrial parks. Specific equipment parameters, load curves, and pipeline topology are shown in Table 1. The data is referenced from actual operating data of an industrial park in the Yangtze River Delta region and anonymized.

[0174] Table 1

[0175] Explanation of waste heat sources: The waste heat of main body A comes from the exhaust gas of the gas turbine; the waste heat of main bodies B and C comes from internal industrial kilns, air compressors and other process equipment, and not only from power generation equipment.

[0176] S5.3 Simulation Result Analysis:

[0177] The effectiveness of this method is verified by comparing the following three schemes:

[0178] Option 1: Traditional centralized dispatching (without distinguishing between heat energy grades);

[0179] Option 2: Differentiate heat energy grades but adopt centralized dispatching;

[0180] Option 3: The method of this invention (differentiating thermal energy grade + distributed optimization).

[0181] Table 2

[0182]

[0183] Results analysis:

[0184] Energy efficiency improvement: The method of this invention improves thermal energy utilization efficiency by about 19 percentage points (absolute value) compared with the centralized scheduling scheme (Scheme 1) that does not distinguish thermal energy grade by matching thermal energy grade and utilizing waste heat across the main body.

[0185] Cost and Emissions Reduction: The overall cost is reduced by approximately 9% compared to Option 1, and carbon emissions are reduced by approximately 17%. The main benefit comes from the cascade utilization of thermal energy, reducing the amount of purchased energy. The above data are simulation calculations; the actual effect is affected by factors such as the amount of waste heat resources in the park and pipeline network conditions, and needs to be evaluated based on specific scenarios.

[0186] Privacy protection: During the solution process, the subjects only exchange coupled variables such as heat transaction volume and green electricity transaction volume, and internal equipment parameters and cost data are not disclosed.

[0187] Solution efficiency: With 3 principals, the distributed solution time is approximately 7 seconds, meeting the day-ahead scheduling time requirements. When the number of principals is expanded to 10, the measured solution time is approximately 45 seconds, still within an acceptable range.

[0188] Preferred embodiments and simulation verification:

[0189] This embodiment takes a three-entity industrial park as an example, and completes the simulation verification of the scheduling strategy according to steps S1 to S5. The scenario setting of the embodiment is as follows: Entity A is a chemical enterprise with high-temperature heat demand; Entity B is a food processing enterprise with medium-temperature heat demand; Entity C is an office area with low-temperature heat demand. The gas turbine of Entity A generates 350°C flue gas, which, after meeting its own 250°C process requirements, generates 180°C waste heat, which is transported to Entity B through the heating network; Entity B generates 70°C waste heat after use, which is transported to Entity C for heating. Specific simulation parameters are shown in Table S5.2, and simulation results are shown in Table 2.

[0190] This embodiment verifies the effectiveness of the method of the present invention in terms of energy efficiency improvement, carbon emission reduction, privacy protection, and computational feasibility. The above performance data comes from the MATLAB / Gurobi simulation environment, and the test system is constructed based on anonymized data from a typical industrial park. The effects in actual engineering applications may vary due to factors such as park size, equipment type, operating conditions, and market environment.

[0191] Example 2

[0192] Figure 5 This is a schematic diagram of the structure of an electronic device provided in Embodiment 2 of the present invention. The electronic device 10 is intended to represent various forms of digital computers, and may also represent various forms of mobile devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the invention described and / or claimed herein.

[0193] like Figure 5 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded into the RAM 13 from storage unit 18. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0194] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0195] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, central processing units (CPUs), graphics processing units (GPUs), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any suitable processors, controllers, microcontrollers, etc. Processor 11 executes the multi-entity energy carbon collaborative scheduling method for campus thermal energy grade stratification matching and full privacy protection described above.

[0196] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0197] The above embodiments are merely illustrative examples and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A multi-entity energy and carbon collaborative scheduling method for hierarchical matching of thermal energy quality and full privacy protection in industrial parks, characterized in that, Includes the following steps: S1: Based on the temperature range, thermal energy is divided into multiple grade levels. For each grade level, the energy quality coefficient based on the second law of thermodynamics is used to convert the actual heat of different grades into equivalent work to characterize their quality value. The energy quality coefficient is determined based on the ratio of the heat source temperature to the ambient temperature. S2: Construct a multi-agent energy and carbon collaborative optimization scheduling model that includes local optimization objective functions, local constraints, and multi-agent global coupling constraints for each agent. In the local optimization objective function, energy efficiency evaluation and cost accounting are characterized by the equivalent work output. The local constraints include energy balance constraints based on actual heat output and heat grade matching constraints based on temperature range and supply-demand temperature difference. The multi-agent global coupling constraints include at least grade-based cross-agent heat trading total balance constraints. S3: The alternating direction multiplier method is adopted, and the coupling variables between each entity and the park collaboration layer are exchanged for iterative solution. Under the premise of protecting the internal data privacy of each entity, the multi-entity energy and carbon collaborative optimization scheduling model is solved to obtain the optimal collaborative scheduling strategy. Among them, the coupling variables include at least the graded heat trading volume between each entity.

2. The multi-entity energy and carbon collaborative scheduling method for hierarchical matching and full privacy protection of park thermal energy quality as described in claim 1, is characterized in that, Step S1 includes: The heat energy grade classification divides heat energy into high temperature grade, medium temperature grade and low temperature grade according to the industrial heat temperature range; The energy-mass coefficient, based on the second law of thermodynamics, is calculated using the following formula to convert the actual heat of different grades into equivalent work to characterize their quality value: , ;in, Energy mass coefficient, Where T is the ambient temperature, H is the heat source temperature, and H is the initial heat. Equivalent work; Furthermore, in the multi-subject energy-carbon collaborative optimization scheduling model described in step S2, the energy balance constraint uses the original heat.

3. The multi-entity energy and carbon collaborative scheduling method for stratified matching of park thermal energy quality and full privacy protection as described in claim 1, is characterized in that... Step S1 further includes: establishing an engineering constraint model for cross-body waste heat transport, wherein the engineering constraint model includes a transport distance constraint for limiting the transport distance, a temperature drop model for calculating temperature loss during transport, a heat exchange efficiency constraint for characterizing the heat exchanger efficiency, and a heat flow-temperature-pressure coupling constraint for relating heat, temperature and flow rate.

4. The multi-entity energy and carbon collaborative scheduling method for stratified matching of park thermal energy quality and full privacy protection as described in claim 1, is characterized in that, In step S2, the local optimization objective function of each entity aims to minimize the comprehensive cost of the entity within a single scheduling cycle. The comprehensive cost includes energy procurement cost, which represents the cost of purchasing electricity and gas; operation and maintenance cost, which represents the cost of operating and maintaining equipment; carbon emission cost, which represents the cost of carbon quota trading; and green transaction cost, which represents the cost of trading green electricity and green certificates.

5. The multi-entity energy and carbon collaborative scheduling method for stratified matching of park thermal energy quality and full privacy protection as described in claim 1, is characterized in that, In step S2, the thermal energy grade matching constraint includes: The actual supply temperature of the heat flow must meet the temperature range corresponding to its grade; The actual supply temperature of the heat flow must be greater than or equal to the sum of the required temperature of the heat load at the receiving end and the preset minimum heat exchange temperature difference; Heat-temperature-flow coupling equations are used to correlate heat, temperature, and flow rate.

6. The multi-entity energy and carbon collaborative scheduling method for stratified matching of park thermal energy quality and full privacy protection as described in claim 5, is characterized in that, The heat-temperature-flow coupling equation is a nonlinear constraint, which is transformed into a linear constraint by a piecewise linearization method. The transformed model is a multi-agent energy and carbon collaborative optimization scheduling model in the form of mixed integer linear programming (MILP).

7. The multi-entity energy and carbon collaborative scheduling method for stratified matching of park thermal energy quality and full privacy protection as described in claim 1, is characterized in that, Step S3 specifically includes: Initialize the number of iterations, the initial values ​​of the coupling variables of each subject, and the Lagrange multipliers; After receiving the coupling variables, each entity independently solves its own local optimization subproblem to update its local runtime plan and coupling variables; Each entity will send the updated coupling variables to the park's collaboration layer; The park collaboration layer calculates the residual of the global coupling constraint based on the coupling variables sent by each subject. If the residual is less than the preset convergence threshold, the iteration is terminated. Otherwise, the Lagrange multipliers are updated based on the residual and the iteration returns to the step of independently solving their respective local optimization subproblems.

8. The multi-entity energy and carbon collaborative scheduling method for stratified matching of park thermal energy quality and full privacy protection as described in claim 7, is characterized in that, Step S3 also includes a convergence acceleration strategy, which includes at least one of the following strategies: an adaptive penalty factor adjustment strategy, which dynamically adjusts the penalty factor based on the ratio of the original residual to the dual residual. The mechanism includes an early stop mechanism, which terminates the iteration early when the cost reduction is less than a preset value; and a hot start strategy, which uses the optimization result of the previous scheduling cycle as the initial solution for the current cycle.

9. The multi-entity energy and carbon collaborative scheduling method for hierarchical matching and full privacy protection of park thermal energy quality as described in claim 1, characterized in that, The method further includes step S4: using a stochastic optimization method to handle the uncertainties of new energy output, load, and market prices, specifically including: A multi-stage stochastic programming framework is adopted. Multiple uncertainty scenarios are obtained by sampling and clustering reduction. The scheduling problem is decomposed into day-ahead and day-intraday stages, and step S3 is called to solve the problem under each uncertainty scenario.

10. The multi-entity energy and carbon collaborative scheduling method for hierarchical matching and full privacy protection of park thermal energy quality as described in claim 9, characterized in that, The multi-stage stochastic programming framework uses stochastic dual dynamic programming as the outer framework and step S3 as the inner solver. The stochastic dual dynamic programming is used to perform cross-scenario coordination optimization and update state variables. Step S3 is used to solve the multi-agent scheduling subproblem in each deterministic scenario. The two are nested and iterated until convergence.

Images