Layered game optimization scheduling and income distribution method for load aggregator

Abstract

The invention belongs to the technical field of power system scheduling, and discloses a load aggregator-oriented hierarchical game optimization scheduling and income distribution method. According to the method, unified aggregation of 10 kV / 35 kV multi-voltage-level users is realized through standardized modeling, and multi-market general bidding parameters are extracted; an upper-layer optimization framework covering multiple markets such as electric energy and demand response is constructed, and a particle swarm algorithm and a commercial solver are adopted for cooperative game solving; taking an upper-layer result as input, and combining a voltage grade difference and performance score recursion mechanism to complete lower-layer user scheduling distribution and income allocation; and through physical feasibility and budget balance verification and credibility, reversely correcting closed-loop cooperation of upper-layer bidding until convergence. According to the method, the market default risk can be reduced, the scheduling robustness and profit potential can be improved, the distribution fairness and the user retention stability are considered, the scheme can be ensured to be on the ground, and the market competitiveness and the operation safety of the load aggregator are comprehensively enhanced.

Description

Technical Field

[0001] This invention relates to the field of power market operation and load resource dispatching technology, specifically a hierarchical game-theoretic optimization dispatching and revenue distribution method for load aggregators. Background Technology

[0002] With the continuous advancement of power system reform and the gradual improvement of the power market trading mechanism, the electricity market, demand response market, and ancillary service markets such as peak shaving and reserve are constantly developing, and the proportion of load-side resources participating in power system dispatch and market transactions is continuously increasing. As an important intermediary connecting distributed users and the power market, load aggregators participate in power market transactions and ancillary services as a whole by aggregating industrial and commercial users, mining farms, and other adjustable load resources, providing new regulatory means for the safe operation of the power system and the consumption of new energy.

[0003] In practical applications, the load resources managed by load aggregators typically exhibit diverse voltage levels, large variations in load size, and inconsistent operating characteristics. For example, 10kV general industrial and commercial users have relatively small load sizes and high regulation flexibility, while large industrial users such as 35kV mines have large load sizes, exhibit significant continuous operation characteristics, and face stricter regulation depth and response constraints. Users at different voltage levels also show significant differences in retail electricity prices, electricity cost structures, and economic incentives for participating in regulation. How to uniformly model and coordinate the scheduling of loads at multiple voltage levels within the same aggregation system is a key engineering problem faced by load aggregators.

[0004] In existing technologies, the optimization model for load aggregators participating in the electricity market typically adopts a hierarchical modeling approach: at the upper level, they participate in the joint clearing of the electricity and ancillary services market based on the aggregated load size and adjustability; at the lower level, they perform performance allocation and revenue settlement for internal users based on the clearing results of the upper level. However, the above hierarchical model generally suffers from the following shortcomings in practical applications.

[0005] On the one hand, the coupling between the upper and lower layer models is relatively weak. Existing methods mostly use the clearing results of the upper layer as a fixed input for the scheduling of the lower layer. However, the actual performance, execution deviations and reliability differences of the lower layer users are usually only used for ex-post assessment or economic penalties. It is difficult to provide effective feedback to the upper layer decision-making in subsequent transactions and bidding processes, which leads to the risk that load aggregators have excessively large bids and insufficient actual executable capabilities in the market.

[0006] On the other hand, the methods for depicting user performance are rather crude. Existing technologies mostly use static credit evaluation, single deviation penalties, or simple performance ratios to evaluate user response, lacking quantitative indicators that can be recursively updated over time and reflect long-term performance reliability. It is also difficult to organically combine performance with scheduling priority, capability boundaries, and revenue distribution mechanisms.

[0007] Furthermore, in terms of internal settlement and revenue distribution among load aggregators, existing methods often focus on meeting short-term performance constraints, and do not adequately consider the comprehensive impact of differences in voltage levels, electricity prices, and performance. This can easily lead to insufficient incentives for some high-voltage level users or users with high performance, thereby affecting the long-term stability of aggregated resources and their enthusiasm for participation.

[0008] In summary, existing load aggregator scheduling and market participation methods still have shortcomings in areas such as unified modeling of loads at multiple voltage levels, quantification of performance reliability, upper and lower layer collaboration, and long-term closed-loop operation mechanisms. There is an urgent need for a technical solution that can tightly couple user performance behavior with upper and lower layer optimization processes and enable load aggregators to stably participate in multi-market transactions while ensuring physical and economic feasibility. Summary of the Invention

[0009] To address the shortcomings of existing technologies, this invention provides a hierarchical game-theoretic optimization scheduling and revenue distribution method for load aggregators, aiming to solve the problems in the background technology.

[0010] To achieve the above objectives, the present invention provides the following technical solution: a hierarchical game-theoretic optimization scheduling and revenue allocation method for load aggregators, comprising the following steps: Step S1: Construct a basic model for users of multiple voltage levels under the jurisdiction of the load aggregator, and uniformly describe the baseline load, adjustability and performance score of each user; Step S2: Based on the basic model, further aggregate to form the equivalent baseline load and flexible interval model at the load aggregator level, and extract unified bidding parameters for joint clearing in multiple markets; Step S3: Based on unified bidding parameters, construct a multi-market joint clearing model covering electricity, demand response, peak shaving and backup services, and form an upper-level optimization framework with the goal of maximizing the total revenue of load aggregators and the constraints of performance capability and flexible range. Step S4: For the multi-market joint clearing model, a multi-agent game mechanism is introduced, and the particle swarm optimization algorithm and commercial solver are used to solve iteratively in collaboration. Finally, the solution converges to the multi-agent game equilibrium state and the upper-level clearing result is output. Step S5: Using the above layer clearing results as input, construct the load aggregator's internal scheduling and revenue distribution model. Combine voltage level differences, recursive updates of performance scores, and deviation penalty mechanisms to complete the allocation of load adjustment instructions for each user, decompose response tasks, and calculate revenue and cost sharing, forming the internal scheduling and revenue distribution results, and outputting a reliable availability feedback signal. Step S6: Based on the internal scheduling and revenue distribution results, establish a physical feasibility constraint and budget balance constraint system, perform consistency verification on the internal scheme, and return to step S5 for adjustment if the constraints are not met. Step S7: Based on the verification results of Step S6, improve the recursive update mechanism of performance score and the feedback mechanism of trusted availability capability. The trusted availability capability is used to correct the fitness function of the upper-level bidding boundary and the optimization algorithm in reverse, forming a closed-loop collaborative operation mode until the convergence condition is met, and the final optimization result is output.

[0011] Furthermore, the specific process of step S1 is as follows: Step S1.1: Define the multi-voltage level aggregated user set: Let the user set under the jurisdiction of the load aggregator be... Represented as: ; This refers to the set of ordinary industrial users with an access voltage level of 10kV; This represents the set of large-load users, such as those in mining areas, with an access voltage level of 35kV; let the user index be... , The total number of users is ; Step S1.2, Unified Modeling of User Baseline Load: Assume users Within the scheduling cycle, the first The baseline load for each time period is ; Step S1.3: Modeling User Adjustability: Based on the baseline load, introduce the user's load adjustment capability under the action of the electricity trading market and dispatch instructions; assume the user... During the period The maximum adjustable down and maximum adjustable up are respectively , Define users in this way During the period Maximum feasible operating power range ; Step S1.4: Establish a user performance scoring model oriented towards revenue distribution: Let the user... During the period The planned regulation power and the actual response power are respectively , Define user During the period Performance deviation Based on the statistical results of users' performance deviations over multiple transaction periods, a user... Performance rating ; Step S1.5: Unify the output format of the basic model: For ordinary industrial users with a voltage level of 10kV or large-load users such as mines with a voltage level of 35kV, a unified representation is provided. .

[0012] Furthermore, the specific process of step S2 is as follows: Step S2.1: Construct the equivalent baseline load model for the load aggregator: Define the load aggregator in the time period. The equivalent baseline load is the weighted sum of the baseline loads of all aggregated users, i.e. , Indicates the load aggregator during the time period The equivalent baseline load; Step S2.2: Construct an aggregated adjustable capacity model considering performance scoring: Define the load aggregator in the time period. Equivalent down-adjustment capability and equivalent up-adjustment capability , ; Step S2.3: Construct a flexible interval model for load aggregators: The load aggregator operates within a specific time period. The flexible operating range is: , Indicates the load aggregator during the time period The equivalent declared load power; Step S2.4: Construct unified bidding parameters for multi-market joint clearing: Abstract the equivalent model of load aggregator into time periods. Standardized set of bidding parameters for multi-market joint clearing .

[0013] Furthermore, the specific process of step S3 is as follows: Step S3.1: Define multi-market participant decision variables: Assume the load aggregator is in the time period Market participation decision variables include , , , ,in, Indicates the load aggregator during the time period The equivalent load power declared in the electricity market. Indicates the load aggregator during the time period Demand response market commitment to load regulation capacity, Indicates the load aggregator during the time period The peak-shaving capacity provided by the peak-shaving ancillary services market Indicates the load aggregator during the time period The standby capacity declared in the standby service market; Step S3.2: Construct the multi-market joint clearing objective function: With the goal of maximizing the total revenue of the load aggregator within the scheduling period, the multi-market joint clearing objective function is as follows: ; In the formula, This represents the total market revenue of the load aggregator during the scheduling period; Indicates the electricity market period Transaction electricity price; Indicates demand response market period Compensation price; Indicates peak-shaving ancillary service market period Service prices; Indicates the standby service market period Capacity and price; Step S3.3: Set load aggregation quotient power balance and flexible range constraints: Equivalent load power balance constraint for load aggregator: ; Flexible interval constraints: ; Step S3.4: Set performance scoring constraints: .

[0014] Furthermore, the specific process of step S4 is as follows: Step S4.1, Multi-agent game modeling and clearing mapping: Let the set of market entities participating in joint clearing be . , Indicates the load aggregator; each entity Through the bid parameter vector Participating in the game, the bidding parameters of the load aggregator are further parameterized into Given all subject bidding parameters Under these conditions, the market operator calls a commercial optimization solver to solve the multi-market joint clearing model in step S3, forming a clearing result vector. This process can be abstracted as a clearing mapping relationship: , Represents the feasible region. This represents the joint clearing objective function. This represents the decision variable to be solved in the multi-market joint clearing model; Step S4.2, Optimal Response Game Theory and Particle Swarm Optimization Algorithm Outer Layer Search: In the game iteration, each agent adopts an iterative optimal response mechanism. In this iteration, when other entities bid parameters When fixed, the main body The optimal response problem is: ; In the formula, Representing the subject In the The bidding parameter vector in the next iteration; Indicates the first In the next iteration, the main body Bidding parameter vectors of other market participants; Representing the subject Payoff function; Representing the subject A set of feasible bidding strategies; Step S4.3: Set the particle swarm optimization algorithm update rules and fitness function: Represent the particle position as a set of bidding parameters for the load aggregator. The particle velocity is Define update rules; fitness function for: , Indicates risk penalty items; This represents the risk weighting coefficient; Step S4.4: Execute the particle swarm optimization algorithm-commercial solver collaborative iterative process.

[0015] Furthermore, the specific process of step S5 is as follows: input the upper-layer clearing results into the lower-layer model, namely the load aggregator's internal scheduling and revenue allocation model, and call the commercial solver to complete the internal execution allocation, revenue allocation, and performance score update, including: Step S5.1: Define the coupling input and resource set from the upper layer to the lower layer: The demand response scale of the upper layer clearing result. , and the amount of the reserve bid , and the amount of the reserve bid Peak shaving winning bid quantity and the distributable revenue of load aggregators As input parameters for the lower-level model, namely the internal scheduling and revenue distribution model of the load aggregator; Indicates the load aggregator during the time period The number of bids awarded in response to demand; Step S5.2: Construct a recursive update model for performance scoring and a deviation penalty model: A recursive update mechanism for performance scoring is adopted; assuming the user... During the period The allocated planned execution volume is The actual execution volume is Define normalization bias , Represents a small constant; based on normalized bias Construct a recursive update formula for performance scoring, expressed as: ; In the formula, Indicates the forgetting factor; Indicates the interval truncation operator; Indicates user During the period Performance rating; Introducing a threshold penalty term , Indicates user During the period Threshold penalty term, Indicates the minimum performance threshold. Indicates the penalty coefficient. Indicates the positive operator; Step S5.3: Constructing Voltage Level Electricity Price Difference Correction and Overall Weighting: Let the voltage level be... voltage level During the period The time-of-use retail electricity price is Define voltage levels During the period Voltage correction factor , This indicates the 10kV voltage level during the time period. Time-of-use retail electricity pricing; further introduce a weighted term for performance differences to construct user... During the period Overall weight , Indicates user The voltage level belongs to the time period Voltage correction factor, Indicates user Voltage level This represents the performance weighting adjustment coefficient. Indicates time period Average performance level; Step S5.4: Construct the internal scheduling and revenue distribution model for the load aggregator: Decision variables include users. During the period Execution volume ,user During the period upper reserve commitment ,user During the period The next reserve commitment and users During the period Internal settlement electricity price Or by user During the period payment amount As equivalent variables; the objective function is: ; In the formula, Indicates user During the period The opportunity cost or negative utility cost function of executing the response; This indicates that the revenue of the load aggregator is retained by weight; This represents the retained revenue item of the load aggregator within its internal mechanism; Step S5.5: Set strong coupling constraints, user individual trust capability boundary constraints, and budget feasibility constraints; Step S5.6: Generate internal settlement price and payment rules: Define users During the period Settlement share In turn, gain users During the period payment amount And reverse the user During the period Internal settlement electricity price ; Step S5.7: Outputting Lower-Level Results and Building a Closed-Loop Feedback Interface: The lower layer outputs internal scheduling and revenue distribution results, including the execution volume of each user. , Reserve Commitment , reserve commitment and internal settlement electricity price and payment amount and update the performance score. The performance score is converted into a reliable availability capability. .

[0016] Furthermore, the specific process of step S6 is as follows: Step S6.1: Define internal execution objects and variables: Let the set of internal users of the load aggregator be... For any user During the scheduling period Internally, the relevant variables for execution include: execution volume. , Reserve Commitment , reserve commitment Maximum adjustable capacity Performance rating and trusted availability ; Step S6.2: Construct an internal execution physical feasibility constraint system; Step S6.3: Construct an internal settlement budget balance constraint system: Assume the user During the period The internal settlement payment amount is Set the budget balance constraint as follows: , This indicates that the upper levels jointly cleared out the total revenue; Step S6.4: Establish a linkage mechanism between budget balancing and performance penalties: when performance scoring... At that time, threshold penalty term Reduce its internal payments; Step S6.5, Execution Consistency and Economic Consistency Verification Output: Physical consistency verification checks whether the execution variables meet the execution result capability composite constraint, capability mutual exclusion constraint and upper-level result consistency constraint; economic consistency verification checks whether the total internal payment amount meets the budget balance constraint; if the verification passes, the final result is output; if it fails, return to step S5 for readjustment.

[0017] Furthermore, the specific process of step S7 is as follows: Step S7.1: Perform recursive updates to the performance score: Assume the user... During the period The allocated planned execution volume is The actual execution volume is Normalization bias is The recursive update formula for performance scoring is: ; Step S7.2, Execute performance threshold determination and trustworthiness mapping: Introduce a minimum performance threshold. ,when The system determines that the user's response reliability is insufficient; a mapping relationship between "performance score" and "trustworthiness" is constructed. , Indicates user During the period Maximum available capacity; Step S7.3: Implement the feedback effect of trusted capabilities on lower-level scheduling and revenue distribution: Trusted availability capabilities are subject to the constraints of steps S5 and S6. Step S7.4: Output trusted and available capabilities from the lower layer. Feedback is provided to correct the upper-level bidding boundaries; Indicates user During the period Trusted availability capability Indicates user During the period The credibility coefficient, Indicates user During the period Maximum available capacity; Step S7.5: Establish a reliable capability feedback mechanism for the upper-level bidding boundary: When the load aggregator constructs the next round of bidding parameters, it will use the adjustable capability boundary at the aggregation level, i.e., the maximum adjustable capability initially based on each user's nominal value. The aggregated result is the initial adjustable capability range for external market bidding, corrected to a reliable and available capability. , Indicates time period A reliable and flexible upper limit that can be used for upper-level bidding; Step S7.6: Achieve a closed-loop collaborative mechanism between upper and lower layers coupled with the intelligent algorithm: Load aggregator bidding parameters Updated by particle swarm optimization algorithm; Step S7.7: Execute closed-loop operation and convergence criteria: A closed-loop operation mode is formed through the cycle of "upper-layer joint clearing - lower-layer execution and allocation - performance score update - trust capability feedback - intelligent algorithm update". When two adjacent iterations satisfy the criteria, the closed-loop operation mode is established. or When the system reaches a stable state, it outputs the final upper-layer clearing result, the lower-layer internal scheduling and revenue distribution result, and the final stable user performance score. , They represent the first , In the round of iteration, time period A reliable and flexible upper limit that can be used for upper-level bidding. Indicates the convergence threshold. , They represent the first , In each iteration, the revenue item of the load aggregator; Step S7.8: Determine whether the convergence condition is met. If not, return to step S7.1 to continue iterating.

[0018] Furthermore, strong coupling constraints, user individual trust capability boundary constraints, and budget feasibility constraints are set, including: Strong coupling constraints are expressed as: ; The boundary constraint of a user's individual trustworthiness is expressed as: , Indicates user During the period The maximum boundary of the upper reserve commitment. Indicates user During the period The maximum boundary of the next reserve commitment; The coupling constraint that "the same capability cannot be sold repeatedly" is expressed as: ; The balance of payments, or budgetary feasibility constraint, is expressed as: , This indicates the percentage reserved by the aggregator.

[0019] Furthermore, an internal physical feasibility constraint system is constructed, including: Set composite constraints on execution result capabilities: ; Set mutual exclusion constraints for capabilities: ; Set consistency constraints for upper-level results: .

[0020] Compared with existing technologies, this invention has the following advantages: It achieves unified aggregation of loads at multiple voltage levels (10kV / 35kV) through standardized modeling, enabling load aggregators to participate efficiently in various electricity market transactions; it dynamically quantifies user reliability through recursive updates of performance scores and a reliable capability mapping mechanism, reducing market default risk at its source; it constructs a closed-loop collaborative system between upper and lower layers, allowing bidding strategies to dynamically adapt to actual execution capabilities, improving dispatch robustness; it integrates multiple market resources for joint game optimization, fully tapping revenue potential; it designs fair allocation rules based on voltage level differences and performance, balancing incentive effectiveness and user retention stability; and it ensures the dispatch scheme is implementable and financially sound through dual verification of physical feasibility and economic consistency, comprehensively enhancing the market competitiveness and operational security of load aggregators. Attached Figure Description

[0021] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0022] like Figure 1 As shown, the present invention provides a technical solution: a hierarchical game-theoretic optimization scheduling and revenue allocation method for load aggregators, comprising the following steps: Step S1: Construct a basic model of users of multiple voltage levels under the jurisdiction of the load aggregator, and uniformly describe the baseline load, adjustability and performance score of each user (to lay a unified data foundation for subsequent aggregation modeling and transaction participation).

[0023] Step S2: Based on the basic model, further aggregate to form the equivalent baseline load and flexible interval model at the load aggregator level, and extract unified bidding parameters that can be directly used for joint clearing in multiple markets (eliminating the impact of voltage level differences on market bidding).

[0024] Step S3: Based on unified bidding parameters, construct a multi-market joint clearing model covering electricity, demand response, peak shaving and backup services, forming an upper-level optimization framework with the goal of maximizing the total revenue of load aggregators and constrained by performance capability and flexible range.

[0025] Step S4: For the multi-market joint clearing model, a multi-agent game mechanism is introduced, and the particle swarm optimization algorithm and commercial solver are used to solve iteratively. Finally, the solution converges to the multi-agent game equilibrium state and the upper-level clearing result is output.

[0026] Step S5: Using the above layer clearing results as input, construct the load aggregator's internal scheduling and revenue distribution model. Combine voltage level differences, recursive updates of performance scores, and deviation penalty mechanisms to complete the allocation of load adjustment instructions for each user, decompose response tasks, and calculate revenue and cost sharing, forming the internal scheduling and revenue distribution results, and outputting a reliable availability feedback signal.

[0027] Step S6: Based on the internal scheduling and revenue distribution results, establish a physical feasibility constraint and budget balance constraint system, perform consistency verification on the internal plan, and if the constraints are not met, return to step S5 for adjustment (to ensure the feasibility of execution and settlement).

[0028] Step S7: Combining the verification results of Step S6, improve the recursive update mechanism of performance score and the feedback mechanism of reliable availability capability. The reliable availability capability is used to correct the upper-level bidding boundary (load aggregator flexible interval model) and the fitness function of the intelligent algorithm in reverse, forming a closed-loop collaborative operation mode until the convergence condition is met, and the final optimization result is output.

[0029] The specific process of step S1 is as follows: Step S1.1: Define the multi-voltage level aggregated user set: Let the user set under the jurisdiction of the load aggregator be... Represented as: ; In the formula, This refers to the set of ordinary industrial users with an access voltage level of 10kV; This represents the set of large-load users, such as those in mining areas, with an access voltage level of 35kV; let the user index be... , The total number of users is .

[0030] Since there are significant differences between 10kV industrial users and 35kV mining users in terms of load scale, operational continuity, regulation depth and response stability, this invention uses a unified parameter system to provide equivalent descriptions of users at different voltage levels, so as to achieve load aggregation across voltage levels.

[0031] Step S1.2, Unified Modeling of User Baseline Load: Assume users Within the scheduling cycle, the first The baseline load for each time period is , , This indicates the number of discrete time periods within the scheduling cycle; the baseline load is statistically modeled based on users' historical electricity consumption data, with 10kV general industrial users focusing on production shifts, load periodicity and seasonality factors, and 35kV mining users focusing on continuous production characteristics and high-load stable operation ranges.

[0032] Step S1.3: Modeling User Adjustability: Based on the baseline load, introduce the user's load adjustment capability under the action of the electricity trading market and dispatch instructions; assume the user... During the period The maximum adjustable down and maximum adjustable up are respectively , Then the user During the period Maximum feasible operating power range for: .

[0033] This model enables the small-scale flexible adjustment capabilities of 10kV industrial users and the large-scale controllable loads of 35kV mining users to be equivalently described within the same aggregation framework.

[0034] Step S1.4: Establish a user performance scoring model oriented towards revenue distribution: Unlike existing technologies that only use performance status for response evaluation, this invention innovatively introduces performance scoring as the core basis for revenue distribution and incentive constraints within the load aggregator; assuming users... During the period The planned regulation power and the actual response power are respectively , Define user During the period Performance deviation Based on the statistical results of users' performance deviations over multiple transaction periods, a user... Performance rating The performance score reflects the long-term reliability of a user's participation in power trading and dispatching instructions. The higher the performance score, the higher the credibility weight of the user in aggregator bidding and revenue distribution.

[0035] Step S1.5: Unify the output format of the basic model: For ordinary industrial users with a voltage level of 10kV or large-load users such as mines with a voltage level of 35kV, a unified representation is provided. This basic model serves as a unified input for load aggregators participating in the electricity trading market, laying the foundation for subsequent steps such as the construction of the aggregation bidding model, revenue calculation, and the profit distribution mechanism based on performance scoring.

[0036] After completing the unified basic modeling of 10kV general industrial users and 35kV mining users, in order to enable load aggregators to participate in the electricity trading market and multiple ancillary service markets as a single entity, it is necessary to further aggregate the distributed user-side models to form an equivalent load model at the aggregator level. This invention, based on the user-side baseline load, adjustability, and performance scoring models, constructs an overall baseline load and flexible interval model for the aggregator, thereby forming unified bidding parameters that can be directly used for joint clearing in multiple markets.

[0037] The specific process of step S2 is as follows: Step S2.1: Construct the equivalent baseline load model for the load aggregator: Define the load aggregator in the time period. The equivalent baseline load is the weighted sum of the baseline loads of all aggregated users, i.e. , Indicates the load aggregator during the time period The equivalent baseline load (kW) is used to eliminate the impact of voltage level differences on market-side modeling, enabling aggregators to participate in the electricity market declaration as "virtual large users".

[0038] Step S2.2: Constructing an aggregated adjustability model considering performance scoring: To avoid the risk to aggregator market performance caused by user performance uncertainty, this invention introduces a performance scoring weighting mechanism in the calculation of aggregated adjustability; defining the load aggregator in time periods The equivalent down-adjustment capability and equivalent up-adjustment capability are respectively: ; In the formula, Indicates the load aggregator during the time period The equivalent downregulation capability; Indicates the load aggregator during the time period The equivalent upward adjustment capability; By weighting performance scores, the influence of reliable users on aggregated bidding capabilities can be reduced, thereby improving the credibility of overall bidding parameters.

[0039] Step S2.3: Construct a flexible interval model for load aggregators: The load aggregator operates within a specific time period. The flexible operating range is: ; In the formula, Indicates the load aggregator during the time period The equivalent declared load power; This flexible range simultaneously covers the small-scale rapid adjustment capability of 10kV industrial users and the large-scale stable adjustment capability of 35kV mining users, providing a unified power boundary description for aggregators to participate in multiple types of markets.

[0040] Step S2.4: Constructing unified bidding parameters for joint clearing across multiple markets: To achieve joint clearing of the electricity market, demand response market, and ancillary services market, the load aggregator equivalent model is abstracted into time periods. Standardized set of bidding parameters for multi-market joint clearing This parameter can be directly mapped to the baseline electricity, interruptible capacity, or adjustable capacity in market declarations, without distinguishing between user voltage levels, thus achieving consistent modeling across markets and voltage levels.

[0041] Through step S2, this invention transforms loads of different voltage levels, such as 10kV and 35kV, into a unified equivalent model for aggregators; it constrains the aggregator bidding capability through performance scoring, thereby improving market performance reliability; and it constructs unified bidding parameters that can be directly used for joint clearing in multiple markets, reducing transaction complexity.

[0042] After obtaining the equivalent baseline load and flexible interval unified bidding parameters at the load aggregator level, in order to maximize the aggregator's revenue in the power trading market and ensure the feasibility and reliability of regulation behavior, this invention further constructs a multi-market joint clearing model that takes into account the power market, demand response market, peak shaving ancillary service market and reserve service market, forming an upper-level optimization decision-making framework for load aggregators to participate in market transactions.

[0043] The specific process of step S3 is as follows: Step S3.1: Define multi-market participant decision variables: Assume the load aggregator is in the time period Market participation decision variables include , , , ,in, Indicates the load aggregator during the time period The equivalent load power (kW) declared in the electricity market. Indicates the load aggregator during the time period The demand response market commitment to load regulation capacity (kW). Indicates the load aggregator during the time period Peak shaving capacity (kW) provided by the peak shaving ancillary services market. Indicates the load aggregator during the time period The standby capacity (kW) declared in the standby service market.

[0044] Step S3.2: Construct the multi-market joint clearing objective function: With the goal of maximizing the total revenue of the load aggregator within the scheduling period, the multi-market joint clearing objective function is as follows: ; In the formula, This represents the total market revenue of the load aggregator during the scheduling period; Indicates the electricity market period Transaction price of electricity (RMB / kWh); Indicates demand response market period Compensation price (RMB / kW); Indicates peak-shaving ancillary service market period Service price (RMB / kW); Indicates the standby service market period Price per kW (RMB) This objective function unifies the returns from different markets into a single optimization framework, enabling collaborative decision-making across multiple markets.

[0045] Step S3.3: Set load aggregation quotient power balance and flexible range constraints: To ensure the enforceability of multi-market reporting results, the equivalent load power balance constraint of the load aggregator must be met: ; Meanwhile, the regulation capacity provided by load aggregators in each market should meet the flexible range constraint: .

[0046] Step S3.4: Setting Performance Scoring Constraints: To ensure the reliable provision of backup service and adjustment capabilities, this invention further introduces performance scoring constraints, coupling backup capacity with the user's overall performance capability. The performance scoring constraints are as follows: ; The introduction of performance scoring naturally weakens the contribution of low-reliability users to backup service capabilities, thereby reducing the market default risk for aggregators caused by individual users defaulting.

[0047] Based on the aforementioned objective function and constraints, a multi-market joint clearing upper-level optimization model is constructed to enable load aggregators to participate in electricity, demand response, peak shaving ancillary services, and reserve services. This model aims to maximize the overall revenue of aggregators, while using flexible ranges and fulfillment capabilities as constraints, to achieve unified optimization decisions regarding the declared capacity in different markets.

[0048] After completing the construction of the upper-level multi-market joint clearing model and the lower-level internal execution and revenue distribution model, in order to characterize the strategic interaction between load aggregators and other market participants in the joint clearing process and improve the solution efficiency and global optimality of the complex two-level model, this invention further introduces a multi-market participant game mechanism to construct a joint clearing iterative solution method based on the collaboration of intelligent algorithms and commercial optimization solvers.

[0049] The specific process of step S4 is as follows: Step S4.1, Multi-agent game modeling and clearing mapping: Let the set of market entities participating in joint clearing be . , Indicates the load aggregator; each entity Through the bid parameter vector ( ) Participating in the game, the bidding parameters of the load aggregator are further parameterized into Given all subject bidding parameters Under these conditions, the market operator calls a commercial optimization solver to solve the multi-market joint clearing model in step S3, forming a clearing result vector. This process can be abstracted as a clearing mapping relationship: , Represents the feasible region. This represents the joint clearing objective function. This represents the decision variable to be solved in the multi-market joint clearing model.

[0050] Step S4.2, Optimal Response Game Theory and Particle Swarm Optimization Algorithm Outer Layer Search: In the game iteration, each agent adopts an iterative optimal response mechanism. In this iteration, when other entities bid parameters When fixed, the main body The optimal response problem is: ; In the formula, Representing the subject In the The bidding parameter vector in the next iteration; Indicates the first In the next iteration, the main body Bidding parameter vectors of other market participants; Representing the subject Payoff function; Representing the subject A set of feasible bidding strategies; this invention introduces a particle swarm optimization (PSO) algorithm in the outer layer to optimize the bidding parameters of the load aggregator. Perform intelligent search.

[0051] Step S4.3: Set the particle swarm optimization algorithm update rules and fitness function: The particle position is represented as a set of bidding parameters for the load aggregator. The particle velocity is The update rules are as follows: ; In the formula, , They represent the first , In this iteration, the particle velocity in the particle swarm optimization algorithm; , They represent the first , In this iteration, the bidding parameters of the load aggregator; Indicates inertia weight; , All represent learning factors; , All represent random numbers; , These represent the individual optimal and global optimal parameters, respectively. fitness function for: ; In the formula, Indicates risk penalty items; This represents the risk weighting coefficient.

[0052] Step S4.4: Execute the particle swarm optimization algorithm - commercial solver cooperative iterative process: Step S4.41: Generate a set of load aggregator bidding parameters using the particle swarm optimization algorithm. ; Step S4.42: Generate the bid parameters for the load aggregator. Mapped to bid vector ; Step S4.43: Call the commercial solver to complete the upper-level joint clearing and obtain the upper-level clearing results, including time periods. Demand response bid volume Time period Backup winning bid quantity Time period Peak shaving winning bid and the distributable revenue of load aggregators ; among which time periods Backup winning bid quantity Including time period The amount of the backup bid and time period The next reserve bid quantity ; The specific process of step S5 is as follows: Input the upper-layer clearing results into the lower-layer model (load aggregator internal scheduling and revenue distribution model), and call the commercial solver to complete the internal execution allocation, revenue distribution, and performance score update, including: Step S5.1: Define the coupling input and resource set from the upper layer to the lower layer: The demand response scale of the upper layer clearing result. , and the amount of the reserve bid , and the amount of the reserve bid Peak shaving winning bid quantity and the distributable revenue of load aggregators (or income in different time periods) ) is used as the parameter input for the lower-level model (the internal scheduling and revenue distribution model of the load aggregator); Indicates the load aggregator during the time period The amount of bids awarded in response to demand.

[0053] Step S5.2: Constructing a recursive update model for performance scoring and a deviation penalty model: To reflect the reliability differences among different users in response execution and to use it for internal scheduling priority and revenue allocation, this invention adopts a recursive update mechanism for performance scoring; assuming users... During the period The allocated planned execution volume is The actual execution volume is Define normalization bias , Represents a small constant; based on normalized bias Construct a recursive update formula for performance scoring, expressed as: ; In the formula, Indicates the forgetting factor; Indicates the interval truncation operator; Indicates user During the period Performance rating; Introducing a threshold penalty term , Indicates user During the period Threshold penalty term, Indicates the minimum performance threshold. Indicates the penalty coefficient. This indicates the positive operator.

[0054] Step S5.3: Constructing Voltage Level Electricity Price Difference Correction and Comprehensive Weighting: To characterize the impact of the difference in retail electricity prices (or agent-purchased electricity prices) between 10kV and 35kV voltage level users on participation incentives, this invention uses the time-of-use electricity price table as a parameter input to construct a voltage correction coefficient; assuming the voltage level... voltage level During the period The time-of-use retail electricity price is Define voltage levels During the period Voltage correction factor , This indicates the 10kV voltage level during the time period. Time-of-use retail electricity pricing; under normal circumstances, 35kV voltage levels during certain time periods Time-of-use retail electricity price Then, the 35kV voltage level during the time period Voltage correction factor This would give 35kV users a higher correction weight, in order to avoid insufficient marginal incentives for their participation in regulation due to the principle that "the higher the voltage level, the lower the electricity price".

[0055] Further introduce weighted terms for performance differences to construct user... During the period Overall weight , Indicates user The voltage level belongs to the time period Voltage correction factor, Indicates user Voltage level This represents the performance weighting adjustment coefficient. Indicates time period Average performance level.

[0056] Step S5.4: Construct the internal scheduling and revenue distribution model for the load aggregator: Decision variables include users. During the period Execution volume ,user During the period upper reserve commitment ,user During the period The next reserve commitment and users During the period Internal settlement electricity price (or by user) During the period payment amount (as equivalent variables); the objective function is: ; In the formula, Indicates user During the period The opportunity cost or negative utility cost function of executing the response; This indicates that the revenue of the load aggregator is retained by weight; This represents the retained revenue item of the load aggregator in its internal mechanism (which can be used to cover management costs, risk reserves, etc.); this objective enables users with high performance and those at the voltage level requiring compensation to obtain a higher revenue weight under the same cost, thereby forming a sustainable incentive.

[0057] Step S5.5: Set strong coupling constraints, user individual trustworthiness boundary constraints, and budget feasibility constraints: To ensure consistency between the lower-level solution and the upper-level clearing result, a strong coupling constraint is set, represented as: ; Set the boundary constraints for individual user trustworthiness, represented as follows: , Indicates user During the period The maximum boundary of the upper reserve commitment. Indicates user During the period The maximum boundary of the next reserve commitment; At the same time, a coupling constraint of "the same capability cannot be sold repeatedly" is introduced, which is expressed as: ; To ensure internal settlement closure and financial feasibility, a balance of payments (budgetary feasibility) constraint is set, represented as: , This indicates the percentage reserved by the aggregator.

[0058] Step S5.6: Generating Internal Settlement Prices and Payment Rules: To generate auditable internal settlement results, this invention adopts a rule-based generation method of "first calculating the share, then reverse-engineering the payment and internal electricity price"; defining users... During the period Settlement share : ; In the formula, Indicates user During the period The overall weight; Indicates user During the period Execution volume; In turn, users During the period payment amount : ; In the formula, Indicates the allocation deduction coefficient; And reverse the user During the period Internal settlement electricity price : ; The above rules guarantee that the total payment amount within any time period will not exceed the distributable income. Users with poor performance will receive lower net payments due to penalties, and users with higher voltage levels can... To obtain reasonable compensation, we can increase participants' enthusiasm for participation.

[0059] Step S5.7, Outputting Lower-Level Results and Building a Closed-Loop Feedback Interface: The lower layer outputs internal scheduling and revenue distribution results (including the execution volume of each user). , Reserve Commitment , reserve commitment and internal settlement electricity price and payment amount ), and update the performance score. The performance score is converted into a reliable availability capability. It is used for feedback on the next round of bidding boundaries at the upper level.

[0060] The specific process of step S6 is as follows: Step S6.1: Define internal execution objects and variables: Let the set of internal users of the load aggregator be... For any user During the scheduling period Internally, the relevant variables for execution include: execution volume. , Reserve Commitment , reserve commitment Maximum adjustable capacity Performance rating and trusted availability .

[0061] Step S6.2: Construct an internal execution physical feasibility constraint system: Set composite constraints on execution result capabilities: ; To prevent the same user from using the same adjustable capacity for both demand response execution and standby commitments simultaneously during the same time period, a capacity mutual exclusion constraint is set: ; The internal scheduling execution volume must be strictly consistent with the upper-level clearing results to ensure that the load aggregator's external fulfillment capability is not weakened. Therefore, upper-level result consistency constraints are set. .

[0062] Step S6.3: Construct an internal settlement budget balance constraint system: On the basis of meeting the feasibility of execution, in order to ensure that the internal settlement of load aggregators does not have a financial deficit, this step constructs a budget balance constraint to constrain and control the total amount of internal payments. Set user During the period The internal settlement payment amount is ; The net amount that a load aggregator pays internally to all users cannot exceed its distributable revenue from the upper-level market. The budget balancing constraint is set as follows: , This indicates that the upper levels jointly cleared out the total revenue.

[0063] Step S6.4: Establish a linkage mechanism between budget balancing and performance penalties: when performance scoring... At that time, threshold penalty term Reduce its internal payments to free up budget space for compensating high-performing users or creating a risk buffer.

[0064] Step S6.5, Execution Consistency and Economic Consistency Verification Output: Physical consistency verification checks execution variables (execution quantity) , Reserve Commitment , reserve commitment Does it satisfy the composite constraint of execution result capability, the mutual exclusion constraint of capability, and the consistency constraint of upper-level result? The economic consistency check examines the total internal payment amount (user payment amount). If the sum of the values ​​satisfies the budget balance constraint, output the final result if the check passes; otherwise, return to step S5 to readjust.

[0065] The specific process of step S7 is as follows: Step S7.1, perform recursive update of performance score: Assume user During the period The allocated planned execution volume is The actual execution volume is Normalization bias is The recursive update formula for performance scoring is: .

[0066] Step S7.2, Execute performance threshold determination and trustworthiness mapping: Introduce a minimum performance threshold. ,when The system determines that the user's response reliability is insufficient; a mapping relationship between "performance score" and "trustworthiness" is constructed. , Indicates user During the period The maximum available capacity.

[0067] Step S7.3: Implement the feedback effect of trusted capability on lower-level scheduling and revenue distribution: Trusted availability capability serves as an important input parameter for the lower-level model (internal scheduling and revenue distribution model of load aggregator), and is subject to the constraints of steps S5 and S6: .

[0068] Step S7.4: Output trusted and available capabilities from the lower layer. Feedback is used to correct the upper-level bidding boundary (flexible range model for load aggregator). Indicates user During the period Trusted availability capability Indicates user During the period The credibility coefficient, Indicates user During the period The maximum available capacity.

[0069] Step S7.5: Establish a reliable capability feedback mechanism for the upper-level bidding boundary: When the load aggregator constructs the next round of bidding parameters, it will include the adjustable capability boundary at the aggregation layer (the load aggregator initially bases this on the maximum adjustable capability nominally held by each user). The aggregated result (the initial adjustable capability range for external market bidding) is corrected to a reliable and available capability. , Indicates time period A reliable and flexible cap that can be used for upper-level bidding.

[0070] Step S7.6: Achieve a closed-loop collaborative mechanism between upper and lower layers coupled with the intelligent algorithm: Load aggregator bidding parameters The fitness function is searched and updated by intelligent algorithms (such as particle swarm optimization), and its fitness function depends on the total income of the upper-level joint clearing. The corresponding risk level (the risk penalty items arising from insufficient reserves, performance deviations, market settlement fluctuations, etc.) is expressed as follows: ; In the formula, The risk penalty item is constructed from factors such as the probability of default and capability uncertainty reflected in the performance score and credibility capability. As the performance score is recursively updated, the credibility capability boundary changes, thereby altering the upper-level clearing results. And the corresponding profit function value, enabling the intelligent algorithm to automatically tend towards a more robust bidding strategy in subsequent iterations.

[0071] Step S7.7: Execute closed-loop operation and convergence criteria: A closed-loop operation mode is formed through the cycle of "upper-layer joint clearing - lower-layer execution and allocation - performance score update - trust capability feedback - intelligent algorithm update". When two adjacent iterations satisfy the criteria, the closed-loop operation mode is established. or When the system reaches a stable state, it outputs the final upper-layer clearing result, the lower-layer internal scheduling and revenue distribution result, and the final stable user performance score. , They represent the first , In the round of iteration, time period A reliable and flexible upper limit that can be used for upper-level bidding. Indicates the convergence threshold. , They represent the first , In the round of iteration, the revenue item of the load aggregator.

[0072] Step S7.8: Determine whether the convergence condition is met. If not, return to step S7.1 to continue iterating.

[0073] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A hierarchical game-theoretic optimization scheduling and revenue distribution method for load aggregators, characterized in that, Includes the following steps: Step S1: Construct a basic model for users of multiple voltage levels under the jurisdiction of the load aggregator, and uniformly describe the baseline load, adjustability and performance score of each user; Step S2: Based on the basic model, further aggregate to form the equivalent baseline load and flexible interval model at the load aggregator level, and extract unified bidding parameters for joint clearing in multiple markets; Step S3: Based on unified bidding parameters, construct a multi-market joint clearing model covering electricity, demand response, peak shaving and backup services, and form an upper-level optimization framework with the goal of maximizing the total revenue of load aggregators and the constraints of performance capability and flexible range. Step S4: For the multi-market joint clearing model, a multi-agent game mechanism is introduced, and the particle swarm optimization algorithm and commercial solver are used to solve iteratively in collaboration. Finally, the solution converges to the multi-agent game equilibrium state and the upper-level clearing result is output. Step S5: Using the above layer clearing results as input, construct the load aggregator's internal scheduling and revenue distribution model. Combine voltage level differences, recursive updates of performance scores, and deviation penalty mechanisms to complete the allocation of load adjustment instructions for each user, decompose response tasks, and calculate revenue and cost sharing, forming the internal scheduling and revenue distribution results, and outputting a reliable availability feedback signal. Step S6: Based on the internal scheduling and revenue distribution results, establish a physical feasibility constraint and budget balance constraint system, perform consistency verification on the internal scheme, and return to step S5 for adjustment if the constraints are not met. Step S7: Based on the verification results of Step S6, improve the recursive update mechanism of performance score and the feedback mechanism of trusted availability capability. The trusted availability capability is used to correct the fitness function of the upper-level bidding boundary and the optimization algorithm in reverse, forming a closed-loop collaborative operation mode until the convergence condition is met, and the final optimization result is output.

2. The hierarchical game-theoretic optimization scheduling and revenue distribution method for load aggregators according to claim 1, characterized in that: The specific process of step S1 is as follows: Step S1.1: Define the multi-voltage level aggregated user set: Let the user set under the jurisdiction of the load aggregator be... Represented as: ; This refers to the set of ordinary industrial users with an access voltage level of 10kV; This represents the set of large-load users, such as those in mining areas, with an access voltage level of 35kV; let the user index be... , The total number of users is ; Step S1.2, Unified Modeling of User Baseline Load: Assume users Within the scheduling cycle, the first The baseline load for each time period is ; Step S1.3: Modeling User Adjustability: Based on the baseline load, introduce the user's load adjustment capability under the action of the electricity trading market and dispatch instructions; assume the user... During the period The maximum adjustable down and maximum adjustable up are respectively , Define users in this way During the period Maximum feasible operating power range ; Step S1.4: Establish a user performance scoring model oriented towards revenue distribution: Let the user... During the period The planned regulation power and the actual response power are respectively , Define user During the period Performance deviation ; Based on the statistical results of user performance deviations over multiple transaction periods, a user... Performance rating ; Step S1.5: Unify the output format of the basic model: For ordinary industrial users with a voltage level of 10kV or large-load users such as mines with a voltage level of 35kV, a unified representation is provided. .

3. The hierarchical game-theoretic optimization scheduling and revenue distribution method for load aggregators according to claim 2, characterized in that: The specific process of step S2 is as follows: Step S2.1: Construct the equivalent baseline load model for the load aggregator: Define the load aggregator in the time period. The equivalent baseline load is the weighted sum of the baseline loads of all aggregated users, i.e. , Indicates the load aggregator during the time period The equivalent baseline load; Step S2.2: Construct an aggregated adjustable capacity model considering performance scoring: Define the load aggregator in the time period. Equivalent down-adjustment capability and equivalent up-adjustment capability , ; Step S2.3: Construct a flexible interval model for load aggregators: The load aggregator operates within a specific time period. The flexible operating range is: , Indicates the load aggregator during the time period The equivalent declared load power; Step S2.4: Construct unified bidding parameters for multi-market joint clearing: Abstract the equivalent model of load aggregator into time periods. Standardized set of bidding parameters for multi-market joint clearing .

4. The hierarchical game-theoretic optimization scheduling and revenue distribution method for load aggregators according to claim 3, characterized in that: The specific process of step S3 is as follows: Step S3.1: Define multi-market participant decision variables: Assume the load aggregator is in the time period Market participation decision variables include , , , ,in, Indicates the load aggregator during the time period The equivalent load power declared in the electricity market. Indicates the load aggregator during the time period Demand response market commitment to load regulation capacity, Indicates the load aggregator during the time period The peak-shaving capacity provided by the peak-shaving ancillary services market Indicates the load aggregator during the time period The standby capacity declared in the standby service market; Step S3.2: Construct the multi-market joint clearing objective function: With the goal of maximizing the total revenue of the load aggregator within the scheduling period, the multi-market joint clearing objective function is as follows: ； In the formula, This represents the total market revenue of the load aggregator during the scheduling period; Indicates the electricity market period Transaction electricity price; Indicates demand response market period Compensation price; Indicates peak-shaving ancillary service market period Service prices; Indicates the standby service market period Capacity and price; Step S3.3: Set load aggregation quotient power balance and flexible range constraints: Equivalent load power balance constraint for load aggregator: ; Flexible interval constraints: ; Step S3.4: Set performance scoring constraints: .

5. The hierarchical game-theoretic optimization scheduling and revenue distribution method for load aggregators according to claim 4, characterized in that: The specific process of step S4 is as follows: Step S4.1, Multi-agent game modeling and clearing mapping: Let the set of market entities participating in joint clearing be . , Indicates the load aggregator; each entity Through the bid parameter vector Participating in the game, the bidding parameters of the load aggregator are further parameterized into Given all subject bidding parameters Under these conditions, the market operator calls a commercial optimization solver to solve the multi-market joint clearing model in step S3, forming a clearing result vector. This process can be abstracted as a clearing mapping relationship: , Represents the feasible region. This represents the joint clearing objective function. This represents the decision variable to be solved in the multi-market joint clearing model; Step S4.2, Optimal Response Game Theory and Particle Swarm Optimization Algorithm Outer Layer Search: In the game iteration, each agent adopts an iterative optimal response mechanism. In this iteration, when other entities bid parameters When fixed, the main body The optimal response problem is: ； In the formula, Representing the subject In the The bidding parameter vector in the next iteration; Indicates the first In the next iteration, the main body Bidding parameter vectors of other market participants; Representing the subject Payoff function; Representing the subject A set of feasible bidding strategies; Step S4.3: Set the particle swarm optimization algorithm update rules and fitness function: Represent the particle position as a set of bidding parameters for the load aggregator. The particle velocity is Define update rules; fitness function for: , Indicates risk penalty items; This represents the risk weighting coefficient; Step S4.4: Execute the particle swarm optimization algorithm-commercial solver collaborative iterative process.

6. The hierarchical game-theoretic optimization scheduling and revenue distribution method for load aggregators according to claim 5, characterized in that: The specific process of step S5 is as follows: Input the upper-layer clearing results into the lower-layer model, namely the load aggregator's internal scheduling and revenue allocation model, and call the commercial solver to complete the internal execution allocation, revenue allocation, and performance score update, including: Step S5.1: Define the coupling input and resource set from the upper layer to the lower layer: The demand response scale of the upper layer clearing result. , and the amount of the reserve bid , and the amount of the reserve bid Peak shaving winning bid quantity and the distributable revenue of load aggregators As input parameters for the lower-level model, namely the internal scheduling and revenue distribution model of the load aggregator; Indicates the load aggregator during the time period The number of bids awarded in response to demand; Step S5.2: Construct a recursive update model for performance scoring and a deviation penalty model: A recursive update mechanism for performance scoring is adopted; assuming the user... During the period The allocated planned execution volume is The actual execution volume is Define normalization bias , Represents a small constant; based on normalized bias Construct a recursive update formula for performance scoring, expressed as: ； In the formula, Indicates the forgetting factor; Indicates the interval truncation operator; Indicates user During the period Performance rating; Introducing a threshold penalty term , Indicates user During the period Threshold penalty term, Indicates the minimum performance threshold. Indicates the penalty coefficient. Indicates the positive operator; Step S5.3: Constructing Voltage Level Electricity Price Difference Correction and Overall Weighting: Let the voltage level be... voltage level During the period The time-of-use retail electricity price is Define voltage levels During the period Voltage correction factor , This indicates the 10kV voltage level during the time period. Time-of-use retail electricity pricing; further introduce a weighted term for performance differences to construct user... During the period Overall weight , Indicates user The voltage level belongs to the time period Voltage correction factor, Indicates user Voltage level This represents the performance weighting adjustment coefficient. Indicates time period Average performance level; Step S5.4: Construct the internal scheduling and revenue distribution model for the load aggregator: Decision variables include users. During the period Execution volume ,user During the period upper reserve commitment ,user During the period The next reserve commitment and users During the period Internal settlement electricity price Or by user During the period payment amount As equivalent variables; the objective function is: ； In the formula, Indicates user During the period The opportunity cost or negative utility cost function of executing the response; This indicates that the revenue of the load aggregator is retained by weight; This represents the retained revenue item of the load aggregator within its internal mechanism; Step S5.5: Set strong coupling constraints, user individual trust capability boundary constraints, and budget feasibility constraints; Step S5.6: Generate internal settlement price and payment rules: Define users During the period Settlement share In turn, gain users During the period payment amount And reverse the user During the period Internal settlement electricity price ; Step S5.7: Outputting Lower-Level Results and Building a Closed-Loop Feedback Interface: The lower layer outputs internal scheduling and revenue distribution results, including the execution volume of each user. , Reserve Commitment , reserve commitment and internal settlement electricity price and payment amount and update the performance score. The performance score is converted into a reliable availability capability. .

7. The hierarchical game-theoretic optimization scheduling and revenue distribution method for load aggregators according to claim 6, characterized in that: The specific process of step S6 is as follows: Step S6.1: Define internal execution objects and variables: Let the set of internal users of the load aggregator be... For any user During the scheduling period Internally, the relevant variables for execution include: execution volume. , Reserve Commitment , reserve commitment Maximum adjustable capacity Performance rating and trusted availability ; Step S6.2: Construct an internal execution physical feasibility constraint system; Step S6.3: Construct an internal settlement budget balance constraint system: Assume the user During the period The internal settlement payment amount is Set the budget balance constraint as follows: , This indicates that the upper levels jointly cleared out the total revenue; Step S6.4: Establish a linkage mechanism between budget balancing and performance penalties: when performance scoring... At that time, threshold penalty term Reduce its internal payments; Step S6.5, Execution Consistency and Economic Consistency Verification Output: Physical consistency verification checks whether the execution variables meet the execution result capability composite constraint, capability mutual exclusion constraint and upper-level result consistency constraint; economic consistency verification checks whether the total internal payment amount meets the budget balance constraint; if the verification passes, the final result is output; if it fails, return to step S5 for readjustment.

8. The hierarchical game-theoretic optimization scheduling and revenue distribution method for load aggregators according to claim 7, characterized in that: The specific process of step S7 is as follows: Step S7.1: Perform recursive updates to the performance score: Assume the user... During the period The allocated planned execution volume is The actual execution volume is Normalization bias is The recursive update formula for performance scoring is: ; Step S7.2, Execute performance threshold determination and trustworthiness mapping: Introduce a minimum performance threshold. ,when The system determines that the user's response reliability is insufficient; a mapping relationship between "performance score" and "trustworthiness" is constructed. , Indicates user During the period Maximum available capacity; Step S7.3: Implement the feedback effect of trusted capabilities on lower-level scheduling and revenue distribution: Trusted availability capabilities are subject to the constraints of steps S5 and S6. Step S7.4: Output trusted and available capabilities from the lower layer. Feedback is provided to correct the upper-level bidding boundaries; Indicates user During the period Trusted availability capability Indicates user During the period The credibility coefficient, Indicates user During the period Maximum available capacity; Step S7.5: Establish a reliable capability feedback mechanism for the upper-level bidding boundary: When the load aggregator constructs the next round of bidding parameters, it will use the adjustable capability boundary at the aggregation level, i.e., the maximum adjustable capability initially based on each user's nominal value. The aggregated result is the initial adjustable capability range for external market bidding, corrected to a reliable and available capability. , Indicates time period A reliable and flexible upper limit that can be used for upper-level bidding; Step S7.6: Achieve a closed-loop collaborative mechanism between upper and lower layers coupled with the intelligent algorithm: Load aggregator bidding parameters Updated by particle swarm optimization algorithm; Step S7.7: Execute closed-loop operation and convergence criteria: A closed-loop operation mode is formed by the cycle of "upper-level joint clearing - lower-level execution and allocation - performance score update - trust capability feedback - intelligent algorithm update". When two adjacent iterations satisfy the criteria, the closed-loop operation mode is formed. or When the system reaches a stable state, it outputs the final upper-layer clearing result, the lower-layer internal scheduling and revenue distribution result, and the final stable user performance score. , They represent the first , In the round of iteration, time period A reliable and flexible upper limit that can be used for upper-level bidding. Indicates the convergence threshold. , They represent the first , In each iteration, the revenue item of the load aggregator; Step S7.8: Determine whether the convergence condition is met. If not, return to step S7.1 to continue iterating.

9. A hierarchical game-theoretic optimization scheduling and revenue distribution method for load aggregators as described in claim 8, characterized in that: Define strong coupling constraints, user individual trust capability boundary constraints, and budget feasibility constraints, including: Strong coupling constraints are expressed as: ; The boundary constraint of a user's individual trustworthiness is expressed as: , Indicates user During the period The maximum boundary of the upper reserve commitment. Indicates user During the period The maximum boundary of the next reserve commitment; The coupling constraint that "the same capability cannot be sold repeatedly" is expressed as: ; The balance of payments, or budgetary feasibility constraint, is expressed as: , This indicates the percentage reserved by the aggregator.

10. A hierarchical game-theoretic optimization scheduling and revenue distribution method for load aggregators as described in claim 9, characterized in that: Construct an internal execution physical feasibility constraint system, including: Set composite constraints on execution result capabilities: ; Set mutual exclusion constraints for capabilities: ; Set consistency constraints for upper-level results: .

Images