Master-slave game and multi-target fusion driven virtual power plant double-layer optimization scheduling method

Abstract

The invention relates to a master-slave game and multi-target fusion driven virtual power plant double-layer optimization scheduling method. The method comprises the following steps: acquiring preprocessed electric and thermal load data sets and typical scene wind and light output data; a double-layer optimization framework of a virtual power plant operator VPPO and an energy supply and demand coordination system ESDS is constructed, the upper layer of the double-layer optimization framework is a virtual power plant operator VPPO model, and the lower layer of the double-layer optimization framework is an energy supply and demand coordination system ESDS multi-target optimization model; and solving the double-layer optimization framework by adopting a hierarchical solving strategy method to obtain an optimization scheduling result. Compared with the prior art, the method has the advantages of improving scheduling flexibility and accuracy and the like.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power plant scheduling, in particular to a master-slave game and multi-objective fusion driven virtual power plant double-layer optimization scheduling method. BACKGROUND

[0002] As a key carrier of aggregating distributed resources to participate in the electricity market, the optimization scheduling strategy of virtual power plant (VPP) is crucial. Traditional scheduling methods often focus on a single objective, such as maximizing economic benefits, which is difficult to fully respond to the increasingly stringent requirements of low-carbon development; while simply pursuing low-carbon nature may significantly increase operating costs, restricting the commercialization of VPP. VPP urgently needs a double-objective optimization mechanism that can effectively coordinate economic efficiency and low-carbon nature. Existing research has shortcomings in dealing with the complex interaction and interest balance of multiple subjects within VPP, making it difficult to simultaneously guarantee the reasonable demands of all parties and the overall optimization target of the system. Therefore, developing an optimization scheduling strategy that can significantly reduce the comprehensive operating cost of VPP, improve economic benefits, maximize the consumption of renewable energy, reduce system carbon emissions, and coordinate the behavior of internal subjects through a scientific mechanism has become an urgent technical demand to promote the large-scale application of VPP and support the clean and low-carbon transformation of energy.

[0003] In the prior art, the optimization scheduling of virtual power plant lacks comprehensive consideration of source-load uncertainty, multi-objective optimization, and master-slave game, and only focuses on one aspect. For example, Chinese Patent No. CN202410499961.5 discloses a virtual power plant multi-objective economic scheduling method and system considering energy efficiency, which only considers maximizing energy utilization efficiency and minimizing operating cost of virtual power plant, without considering source-load uncertainty and master-slave game. Chinese Patent No. CN202310262511.X discloses a virtual power plant multi-objective optimization scheduling method, which considers a reward-punishment ladder type carbon trading mechanism and double-objective optimization, but also does not consider source-load uncertainty and master-slave game. Existing technologies mostly use single or local optimization ideas, do not build an interest interaction framework for multiple subjects within virtual power plant, and do not comprehensively consider complex scenarios such as source-load uncertainty and market price fluctuations, resulting in poor scheduling flexibility and precision. SUMMARY

[0004] The purpose of the present application is to improve the flexibility and precision of scheduling by providing a master-slave game and multi-objective fusion driven virtual power plant double-layer optimization scheduling method, device and medium.

[0005] The purpose of the present application can be achieved by the following technical solutions: A master-slave game and multi-objective fusion driven virtual power plant double-layer optimization scheduling method, the method comprising: acquiring pre-processed electrical and thermal load data sets and typical scenario wind and light output data; A double-layer multi-objective optimization framework of a virtual power plant operator (VPPO) and an energy supply and demand coordination system (ESDS) is constructed, the double-layer multi-objective optimization framework has a VPPO model as an upper layer and an ESDS multi-objective optimization model as a lower layer, the ESDS multi-objective optimization model includes economic and environmental objectives and corresponding weights, and the VPPO model includes electrical and thermal output powers of the energy supply side in the ESDS after t-period optimization scheduling and electrical and thermal powers of the user side after demand response of t-period loads; The double-layer optimization framework is solved by using a hierarchical solution strategy to obtain an optimization scheduling result.

[0006] Further, the VPPO model is as follows:

[0007]

[0008] wherein, represents a target function of the VPPO model, and T is a number of time periods in a scheduling cycle; represents a revenue of the VPPO from selling energy to the user side in t period; represents an energy supply income of the ESDS in t period; represents a cost of purchasing energy by the VPPO from the energy supply side; represents an electricity transaction amount of the VPPO with a power grid, represents a heat load interruption penalty cost of the VPPO due to insufficient heat supply to the user side, represents a cost of purchasing energy by the user, represents an energy supply income of the ESDS in t period, , respectively represent electrical and thermal powers of the user side after demand response in t period; , respectively represent electrical and thermal output powers of the energy supply side in the ESDS after t-period optimization scheduling; is a transaction amount of the VPPO with the power grid, is a heat load interruption supply amount in t period; is a heat interruption penalty coefficient; , respectively represent purchasing and selling electricity prices of the VPPO in t period, , respectively represent purchasing and selling heat prices of the VPPO in t period; , respectively represent time-of-use electricity prices and on-grid electricity prices of the power grid in t period.

[0009] Further, the electric and heat output power of the energy supply side in the ESDS after the t-period optimization scheduling is:

[0010] wherein, , are the electric and heat output power of the energy supply side in the ESDS after the t-period optimization scheduling, respectively, is the efficiency of the heat exchanger, is the t-period photovoltaic output; is the t-period wind-solar output; is the t-period gas turbine output; is the t-period electric boiler power consumption, is the t-period heat power recovered by the waste heat boiler; is the t-period gas boiler output heat power; is the t-period electric boiler output heat power.

[0011] Further, the electric and heat power of the user side after the t-period user side load participates in the demand response is: Further, the multi-objective optimization model of the energy supply and demand coordination system ESDS is:

[0012] wherein, , are the weight of the economic target and the environmental target, respectively, and are the economic target and the environmental target, respectively.

[0013] Further, the economic target is:

[0014] wherein, represents the t-period user utility function, i.e., the user energy use satisfaction, is the total fuel cost of the energy supply side, is the total fuel cost of the user side.

[0015] Further, the environmental target is:

[0016] wherein, is the carbon dioxide treatment cost; , are the carbon emission coefficients of the gas turbine and the gas boiler, respectively, represents the t-period user utility function, i.e., the user energy use satisfaction.

[0017] Further, the user energy use satisfaction is:

[0018] wherein, 、 is a preference coefficient of the user for consuming electric energy, 、 is a preference coefficient of the user for consuming thermal energy.

[0019] Compared with the prior art, the present application has the following beneficial effects: The present application adopts multi-objective optimization, while considering the economic and environmental objectives of the energy supply and demand coordination system. The prior art often focuses on a single objective. The present application balances different dimensional demands through multi-objective collaborative optimization, helps the sustainable operation of the virtual power plant, adapts to the pursuit of comprehensive benefits in energy transformation, and improves its adaptability and competitiveness in the electricity market and energy system.

[0020] The present application combines master-slave game and multi-objective optimization, so that the dispatching strategy can dynamically respond to multiple complex scenarios, such as random fluctuations in new energy output, variable load demand, and fluctuations in electricity market prices. In the face of these scenarios, the adaptability and accuracy of the prior art dispatching are not good enough. With the help of master decision adjustment and multi-objective balance under the game mechanism, the present application can more accurately adapt to scenario changes, ensure stable and efficient operation of the virtual power plant, and enhance its anti-disturbance ability and flexible regulation level.

[0021] The present application introduces master-slave game to build a game interaction framework between different subjects in the virtual power plant. Compared with the single or local optimization dispatching of the prior art, the present application can simulate the interests and decision logic of each subject, so that the dispatching strategy is more systematic under the coordination of multiple subjects, realizes overall optimization, improves the scientificity and effectiveness of resource allocation of the virtual power plant, and makes various resources accurately cooperate under the game mechanism to tap greater operation value. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 is a flowchart of the present application; Figure 2 is a master-slave game architecture of the virtual power plant; Figure 3 is a typical scenario of wind and light output and electric and thermal load; Figure 4 is a hierarchical solution strategy; Figure 5 is an electric energy optimization dispatching result; Figure 6 is a thermal energy optimization dispatching result; Figure 7 is a purchase and sale electricity price; Figure 8 is a purchase and sale heat price; Figure 9 is a source and load resource classification and aggregation framework. Detailed Implementation

[0023] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0024] Example 1: This invention addresses the shortcomings in balancing the complex interactions and interests of multiple stakeholders within a Virtual Power Plant (VPP), as well as the difficulty in simultaneously ensuring system economy and low carbon emissions. It proposes a two-layer optimization scheduling method for VPPs driven by master-slave game theory and multi-objective fusion. This method includes: first, collecting and processing data to obtain a time-series dataset on the source-load side; quantifying wind and solar uncertainties using scenario generation and reduction methods; and establishing operating cost models, energy efficiency models, and operating constraint models for units within the VPP. Second, based on a centralized-distributed control architecture, a source-load resource classification and aggregation framework (such as...) is proposed. Figure 9 As shown, the system categorizes and aggregates resources on both the source and load sides, establishing separate supply and demand aggregation models. On the demand side, user participation in demand response and energy satisfaction are considered. The proposed Energy Supply and Demand Coordination System (ESDS) integrates the supply and demand sides. Then, master-slave game theory and multi-objective optimization are introduced to construct a master-slave game architecture between the Virtual Power Plant Operator (VPPO) and the ESDS. A two-layer optimization model for the virtual power plant is established, with the upper layer being the VPPO model and the lower layer being the ESDS multi-objective optimization model. The dynamic impact of the pricing strategy of the upper-layer VPPO on the multi-objective weights in the lower-layer ESDS is considered. A hierarchical solution strategy is designed for this two-layer optimization model. The upper layer uses an improved genetic algorithm (GA), while the lower layer uses a combination of particle swarm optimization (MOPSO), entropy weight method (EWM), and quadratic programming (QP). Finally, load-side time-series data and quantized wind and solar power output data are input into the two-layer optimization model, and the optimization results are calculated using the hierarchical solution strategy. Compared with existing technologies, this invention helps to promote detailed research on complex motions between multiple entities, while taking into account both system economy and low carbon emissions.

[0025] On one hand, this invention discloses a two-layer optimization scheduling method for virtual power plants driven by master-slave game theory and multi-objective fusion, comprising the following steps: Step S1: Collect historical data on wind and solar power output on the source side and electrical and thermal loads on the load side of the virtual power plant, and clean the data, including removing outliers and compensating for missing values, to form a high-quality input dataset. Step S2: Based on the preprocessed wind and solar data dataset, scene generation and reduction techniques are used to quantify the uncertainty of wind and solar power, obtain the wind and solar power output of typical scenes and the corresponding scene probabilities, and provide data input for the subsequent optimization scheduling model. Step S3: Establish accurate operating cost models, energy efficiency models, and operating constraint models for controllable energy units such as gas turbines, gas boilers, and electric boilers in the virtual power plant; Step S4: Establish a demand-side load dispatchable potential model to characterize the adjustable potential and behavioral characteristics of user load under time-of-use pricing, and measure user energy satisfaction. Step S5: Based on the models established in steps S3 and S4, aggregate the distributed energy and adjustable load resources within the virtual power plant to construct an energy supply and demand coordination system entity with bidirectional flexible interaction capabilities. Step S6: Introduce master-slave game theory and multi-objective optimization to build a master-slave game architecture between Virtual Power Plant Operator (VPPO) and ESDS, and establish a virtual power plant two-layer optimization model with the VPPO model at the upper layer and the ESDS multi-objective optimization model at the lower layer. Step S7: Perform mathematical analysis on the constructed master-slave game model to verify the existence and characteristics of the equilibrium solution; Step S8: For the complex structure of the two-layer optimization model, design an efficient layered solution strategy to achieve collaborative solution between the upper and lower layers; Step S9: Input the preprocessed electricity and heat load datasets and the typical scenario wind and solar power output data generated in step S2 into the two-layer optimization model, and obtain the optimized scheduling result through the solution strategy of step S8.

[0026] Further, step S1 includes the following steps: Step S1.1: Obtain hourly output time-series data of wind farms and photovoltaic power plants, and extract electricity and heat load data from the user's energy management system, with the time resolution consistent with that of the source side.

[0027] Step S1.2: Perform data cleaning on the collected data. Data cleaning includes outlier detection, correction, and missing value compensation. For data with fewer than 3 consecutive missing values, linear interpolation is used to fill in the missing values ​​using data from adjacent time periods. For data with a large number of missing values, a moving average is used to repair the missing segments using historical data under the same operating conditions.

[0028] Further, step S2 includes the following steps: Step S2.1: Based on the preprocessed wind and solar power output data, generate a historical dataset of wind and solar power output for 24 time periods according to the scheduling time period; Step S2.2: For the dataset of each time period, construct the probability density function of each time period using the nonparametric kernel density estimation method, and calculate the cumulative probability distribution function of each time period; Step S2.3: Use Latin hypercube sampling to generate the initial set of wind and solar power output scenarios and the initial probabilities; Step S2.4: Obtain typical wind and solar power output scenarios and their probability distributions using a scenario reduction method based on probabilistic distance and k-means.

[0029] Furthermore, the unit model established in step S3 within the virtual power plant system includes the operating cost model, energy efficiency model, and operating constraint model for each unit within the system. The operating constraints include upper and lower limits of unit output, ramp-up constraints, etc. The equipment within the system includes gas turbines, gas boilers, electric boilers, energy storage equipment, etc.

[0030] Further, step S4 includes the following steps: Step S4.1: Model the dispatchable potential of electrical and thermal loads.

[0031] Electric load dispatchable potential model:

[0032] in The original electrical load during time period t, For a fixed electrical load during time period t, This represents the maximum transferable electrical load during time period t.

[0033] Heat load dispatchable potential model:

[0034] in The original heat load for time period t. For a fixed heat load during time period t, This represents the maximum transferable heat load during time period t.

[0035] Step S4.2: Use a quadratic user utility function to measure energy satisfaction.

[0036] User energy satisfaction can be described as:

[0037] in , This represents the user's preference coefficient for energy consumption. , This represents the user's preference coefficient for consuming heat energy.

[0038] Furthermore, step S5 aggregates distributed energy resources and adjustable load resources within the virtual power plant to establish a supply-side aggregation model and a demand-side aggregation model, which can be expressed by the following formula: The energy supply-side aggregation model is expressed as:

[0039] in , The electrical and thermal output power of the power supply side during time period t; This refers to the efficiency of the heat exchanger. Provide photovoltaic power output for period t; Contribute to the scenery during period t; For the gas turbine output during time period t; The power consumption of the electric boiler during time period t. The heat power generated by the waste heat boiler during time period t; The output thermal power of the gas-fired boiler during time period t; The output heat power of the electric boiler during time period t.

[0040] The demand-side aggregation model is represented as:

[0041] in The electrical load on the demand side after the demand response in time period t; The original electrical load during time period t; Let t be the response quantity of transferable electrical loads participating in demand response during time period t. A value greater than 0 indicates that this portion of the charge has increased to the current time period. A value less than 0 indicates that this portion of the charge has decreased since the current time period. , Let t be the charging and discharging power of the electrical energy storage during time period t. The heat load on the demand side after the demand response in time period t; The original heat load for time period t; The response amount of the heat load that can be reduced during time period t; The heat storage power during time period t represents the heat storage power, where a value greater than 0 indicates the heat storage power and a value less than 0 indicates the heat release power.

[0042] Furthermore, the virtual power plant two-layer optimization framework established in step S6 includes the following steps: Step S6.1: Based on step S5, introduce master-slave game theory to construct a master-slave game architecture between the virtual power plant operator and the energy supply and demand coordination system. Step S6.2: Model the upper-level virtual power plant operator, who acts as the leader in the master-slave game; Virtual power plant operator model:

[0043]

[0044] Where T is the number of time periods in a scheduling cycle; This represents the revenue that a Virtual Power Plant Operator (VPPO) receives from selling energy to users during time period t. This represents the energy supply revenue of the Energy Supply and Demand Coordination System (ESDS) during time period t; This represents the cost of VPPO purchasing energy from the energy supply side; This represents the electricity transaction amount between VPPO and the power grid. A value greater than 0 indicates the cost of purchasing electricity from the grid, while a value less than 0 indicates revenue from surplus electricity sold to the grid. This indicates the heat load interruption penalty cost incurred by the VPPO due to insufficient heating supply to the user side. , These represent the electrical and thermal power of the user-side load after participating in demand response during time period t; , These represent the electrical and thermal output power of the ESDS after optimized scheduling in time period t; This represents the transaction volume between VPPO and the power grid. A value greater than 0 indicates the power purchased from the power grid, while a value less than 0 indicates the surplus power supplied to the grid. The heat load interruption supply during period t; This is the penalty coefficient for heating interruption; , These are the purchase and sale prices of VPPO for time period t, respectively. , These are the purchase and sale prices of VPPO for time period t, respectively. , These represent the time-of-use electricity price and the grid connection price for the power grid during time period t, respectively.

[0045] Step S6.3: Introduce multi-objective optimization to model the lower-level energy supply and demand coordination system as a follower; The objective function of the energy supply and demand coordination system model includes environmental and economic objectives, as shown below: Environmental objectives:

[0046] in Cost of carbon dioxide treatment; , These are the carbon emission coefficients for gas turbines and gas boilers, respectively. Let t represent the utility function of users during time period t, i.e., user energy satisfaction.

[0047] Economic objectives:

[0048] The two objective functions of the lower-level energy supply and demand coordination system are linearly weighted and expressed as follows:

[0049] in , The weights are assigned to economic and environmental objectives, respectively.

[0050] Further, step S7 proves that an equilibrium solution exists between the Virtual Power Plant Operator (VPPO) and the Energy Supply and Demand Coordination System (ESDS) under the master-slave game. The equilibrium solution... Neither VPPO nor ESDS benefits from unilaterally changing the strategy, as can be expressed by the following formula:

[0051] in The objective function for VPPO; This is the overall objective function for ESDS; The optimal strategy under the equilibrium state of VPPO; The optimal strategy under the equilibrium state of ESDS; , The strategies are VPPO and ESDS, respectively.

[0052] Furthermore, in step S8, a hierarchical solution strategy is designed for the two-layer optimization model. An improved genetic algorithm is used to solve the upper-layer VPPO model, while a quadratic programming, entropy weight method, and particle swarm optimization algorithm are used to solve the lower-layer ESDS multi-objective optimization model.

[0053] Furthermore, in step S9, the preprocessed demand-side electricity and heat load data and the typical scenario wind and solar power output data generated in step S2 are input into the virtual power plant two-layer optimization model established in step S6, and the above-mentioned hierarchical solution strategy is used to obtain the optimized scheduling result.

[0054] Secondly, the present invention provides a computer device comprising a processor, a memory, a network interface, and other core hardware components. The memory stores an instruction set, the network interface enables data interaction with various devices within the virtual power plant, and the processor executes the instructions stored in the memory, thereby driving the device to run the aforementioned master-slave game theory and multi-objective fusion-driven virtual power plant two-layer optimization scheduling method.

[0055] Compared with the prior art, the present invention has the following beneficial effects: (1) The scheduling strategy is more systematic and collaborative: This invention introduces a master-slave game to construct a game interaction framework among different subjects in the virtual power plant. Compared with the single or local optimization scheduling of existing technologies, it can simulate the interests and decision-making logic of each subject, making the scheduling strategy more systematic under the collaboration of multiple subjects, achieving overall optimization, improving the scientificity and effectiveness of virtual power plant resource allocation, and enabling various resources to cooperate accurately under the game mechanism to explore greater operational value.

[0056] (2) Multi-objective optimization with comprehensive consideration: This invention adopts multi-objective optimization, taking into account both the economic and environmental objectives of the energy supply and demand coordination system. Existing technologies often focus on a single objective. This invention, through multi-objective collaborative optimization, can balance the needs of different dimensions, help the virtual power plant operate sustainably, adapt to the pursuit of comprehensive benefits in the energy transition, and enhance its adaptability and competitiveness in the electricity market and energy system.

[0057] (3) Enhanced ability to accurately respond to complex scenarios: This invention combines master-slave game theory with multi-objective optimization, enabling the scheduling strategy to dynamically respond to diverse and complex scenarios, such as random fluctuations in renewable energy output, variable load demand, and fluctuations in electricity market prices. In the face of these scenarios, existing technologies lack adaptability and accuracy in scheduling. This invention, through master-slave decision-making adjustments and multi-objective balancing under a game theory mechanism, can more accurately adapt to scenario changes, ensuring the stable and efficient operation of the virtual power plant and enhancing its anti-interference capability and flexible control level.

[0058] This embodiment provides a two-layer optimization scheduling method for virtual power plants driven by master-slave game theory and multi-objective fusion, such as Figure 1 As shown, firstly, data is collected and preprocessed. Typical output data is obtained from the preprocessed wind and solar data using a scene generation and reduction method. Then, resources within the virtual power plant are modeled, classified, and aggregated to establish a supply-demand aggregation model. Finally, a master-slave game theory and multi-objective optimization are introduced to establish a two-layer optimization model for the virtual power plant, with a VPPO model at the upper layer and an ESDS multi-objective optimization model at the lower layer. A hierarchical solution strategy is proposed for this model. The optimization results show that the scheduling strategy provided in this embodiment is economical and low-carbon.

[0059] The specific steps are as follows: Step S1: Collect historical data from the source and load sides, including historical data on wind and solar power output and electricity and heat loads. Preprocess the data, including removing outliers and compensating for missing values. Step S2: Based on the wind and solar power output data obtained from the above processing, the uncertainty of wind and solar power is quantified by the scene generation and reduction method to obtain the wind and solar power output of typical scenes and the corresponding scene probabilities, providing data input for the subsequent optimization scheduling model. Step S3: Establish accurate operating cost models, energy efficiency models, and operating constraint models for the controllable energy units such as gas turbines, gas boilers, and electric boilers included in the virtual power plant; 1. Gas turbine unit: The fuel cost of gas turbines and gas boilers has a quadratic functional relationship with their output power, which can be defined as follows:

[0060] in , The fuel costs for gas turbines and gas boilers are respectively. This represents the total fuel cost on the energy supply side. , , This represents the cost coefficient of the gas turbine. , , This represents the cost coefficient for gas-fired boilers.

[0061] The output of the gas turbine and gas boiler on the power supply side must meet the rated power constraint:

[0062] in , These are the rated power of the gas generator and the gas boiler, respectively.

[0063] 2. Wind turbine model: The output power of a wind turbine generator is determined by the wind speed, which can be described by a mathematical model as follows:

[0064] in The output power of the wind turbine cluster during time period t; This refers to the rated output power of a single wind turbine generator set. The number of wind turbine units; Real-time wind speed; , , These are the cut-in wind speed, rated wind speed, and cut-out wind speed, respectively.

[0065] The power output constraint of the wind turbine can be expressed as:

[0066] in The maximum output power of the wind turbine cluster during time period t; 3. Photovoltaic Model: Photovoltaic power generation is mainly affected by the actual area of ​​the photovoltaic array receiving sunlight and the intensity of sunlight irradiance when the photovoltaic system is put into operation, which can be described by a mathematical model as follows:

[0067] in Let t be the output power of the photovoltaic system during time period t; The actual area of ​​the photovoltaic array receiving sunlight during time period t; The solar irradiance in the region where the photovoltaic system is located during time period t; This refers to the maximum power point tracking efficiency of the photovoltaic system. The power generation efficiency of the inverter; The ambient temperature during time period t.

[0068] Photovoltaic output constraints can be expressed as:

[0069] in This represents the maximum output power.

[0070] 4. Energy storage model: Energy storage includes electrical energy storage and thermal energy storage, where the electrical energy storage model can be described as follows:

[0071] in To store energy for time period t The energy stored in the t+1 time period is the energy stored in the electric energy storage system after a period of charging or discharging from the t time period. , These represent the charging and discharging power of the energy storage system during time period t; , and These are the self-loss rate, charging efficiency, and discharging efficiency of the energy storage system, respectively.

[0072] Energy storage constraints include state constraints, lifetime constraints, and safe capacity constraints, which are expressed as follows:

[0073] in For initial energy storage, This is the maximum capacity of electrical energy storage; The state of charge of the stored electrical energy during time period t. , These are the minimum and maximum states of charge allowed under safe conditions, respectively. , These are the maximum charging power and the maximum discharging power, respectively. , These are the charging and discharging flags for time period t, respectively. This is the sum of the number of charge-discharge cycles per average scheduling cycle under normal operating life. , These represent the upper and lower limits of ramp speed under charging and discharging conditions, respectively.

[0074] The thermal energy storage model can be described as follows:

[0075] in This represents the energy stored during time period t. The energy stored in the t+1 time period is the energy stored by the thermal energy storage system after a period of heat storage or release from the t time period. The stored and released heat power of the thermal energy storage system during time period t, where a value greater than 0 indicates stored heat and a value less than 0 indicates released heat. , and These are the self-loss rate, heat storage efficiency, and heat release efficiency of the thermal energy storage system, respectively.

[0076] The safety capacity constraints, state of charge constraints, and ramping constraints for thermal energy storage are as follows:

[0077] In the formula For initial heat storage, , These represent the upper and lower limits of heat storage under safe conditions; , These are the upper limits for heat storage and heat release, respectively. , These are the upper and lower limits for heat storage and heat release, respectively.

[0078] Step S4: Establish a virtual power plant demand-side dispatchable potential model and measure user energy satisfaction, including the following steps: Step S4.1: Model the dispatchable potential of electrical and thermal loads.

[0079] Electric load dispatchable potential model:

[0080] in The original electrical load during time period t, For a fixed electrical load during time period t, This represents the maximum transferable electrical load during time period t.

[0081] Heat load dispatchable potential model:

[0082] in The original heat load for time period t. For a fixed heat load during time period t, This represents the maximum transferable heat load during time period t.

[0083] Step S4.2: Use a quadratic user utility function to measure energy satisfaction.

[0084] User energy satisfaction can be described as:

[0085] in , This represents the user's preference coefficient for energy consumption. , This represents the user's preference coefficient for consuming heat energy.

[0086] Step S5: Aggregate distributed energy and adjustable load resources within the virtual power plant, establish a supply-side aggregation model and a demand-side aggregation model, and construct an energy supply and demand coordination system entity with bidirectional flexible interaction capabilities. The energy supply-side aggregation model is expressed as:

[0087] in , The electrical and thermal output power on the energy supply side of the Energy Supply and Demand Coordination System (ESDS) during time period t; This refers to the efficiency of the heat exchanger.

[0088] Energy supply revenue on the supply side comes from energy purchases by virtual power plant operators, which can be described as follows:

[0089] in This represents the energy revenue of EDS during time period t.

[0090] The demand-side aggregation model includes a demand response model, where the electricity load demand response model can be expressed as:

[0091] in The original electrical load during time period t; Let t be the response quantity of transferable electrical loads participating in demand response during time period t. A value greater than 0 indicates that this portion of the charge has increased to the current time period. A value less than 0 indicates that this portion of the charge has decreased since the current time period.

[0092] The following constraints must be met for electrical loads to participate in demand response:

[0093] in , These are the upper and lower limits of transferable electrical load, respectively.

[0094] The heat load demand response model can be expressed as:

[0095] in The original heat load for time period t; The response amount of the heat load that can be reduced during time period t.

[0096] The reduction of heat load must meet the upper limit constraint:

[0097] in This represents the upper limit of the load that can be reduced during time period t.

[0098] The energy cost on the user side includes the cost of purchasing energy and the cost of storing energy, which can be expressed as:

[0099] in The energy cost after the load participates in demand response during time period t; , These are the user's energy purchase cost and energy storage cost, respectively; , These are the charging and discharging cost coefficients for electrical energy storage and thermal energy storage, respectively.

[0100] Step S6: Introduce master-slave game theory and multi-objective optimization, build a master-slave game architecture between Virtual Power Plant Operator (VPPO) and ESDS, and establish a virtual power plant two-layer optimization model with the upper layer being the VPPO model and the lower layer being the ESDS multi-objective optimization model, including the following steps; Step S6.1: Based on step S5, introduce master-slave game theory to construct a master-slave game architecture between the virtual power plant operator and the energy supply and demand coordination system. Step S6.2: Model the upper-level virtual power plant operator, who acts as the leader in the master-slave game; Virtual power plant operator model:

[0101]

[0102] Where T is the number of time periods in a scheduling cycle; This represents the revenue that a Virtual Power Plant Operator (VPPO) receives from selling energy to users during time period t. This represents the energy supply revenue of the Energy Supply and Demand Coordination System (ESDS) during time period t; This represents the cost of VPPO purchasing energy from the energy supply side; This represents the electricity transaction amount between VPPO and the power grid. A value greater than 0 indicates the cost of purchasing electricity from the grid, while a value less than 0 indicates revenue from surplus electricity sold to the grid. This indicates the heat load interruption penalty cost incurred by the VPPO due to insufficient heating supply to the user side. , These represent the electrical and thermal power of the user-side load after participating in demand response during time period t; , These represent the electrical and thermal output power on the power supply side of the ESDS after optimized scheduling in time period t; This represents the transaction volume between VPPO and the power grid. A value greater than 0 indicates the power purchased from the power grid, while a value less than 0 indicates the surplus power supplied to the grid. The heat load interruption supply during period t; This is the penalty coefficient for heating interruption; , These are the purchase and sale prices of VPPO for time period t, respectively. , These are the purchase and sale prices of VPPO for time period t, respectively. , These represent the time-of-use electricity price and the grid connection price for the power grid during time period t, respectively.

[0103] Virtual power plant operators' energy pricing needs to meet the following constraints:

[0104] in , These are the upper and lower limits of VPPO's purchase and sale prices for heat, respectively.

[0105] In addition, the electricity and heat sales prices of virtual power plant operators must also meet the average price constraint, which can be expressed as:

[0106] in , These are the upper limits for average electricity and heat sales prices, respectively.

[0107] Energy trading between a virtual power plant operator and an external power grid is represented as follows:

[0108] in The transaction volume between the virtual power plant operator and the external power grid during time period t.

[0109] When the virtual power plant operator provides insufficient heat, it will cause a heat outage, which can be represented as:

[0110] in The heating supply is interrupted during time period t.

[0111] Interruptions to heating supply must be kept within safe limits, which can be represented as:

[0112] in This represents the maximum amount of heating interruption under safe conditions during time period t.

[0113] Step S6.3: Introduce multi-objective optimization to model the lower-level energy supply and demand coordination system as a follower; The objective function of the energy supply and demand coordination system model includes environmental and economic objectives, as shown below: Environmental objectives:

[0114] in Cost of carbon dioxide treatment; , These are the carbon emission coefficients for gas turbines and gas boilers, respectively. Let t represent the utility function of users during time period t, i.e., user energy satisfaction.

[0115] Economic objectives:

[0116] The two objective functions of the lower-level energy supply and demand coordination system are linearly weighted and expressed as follows:

[0117] Step S7: Verify the strategic equilibrium between the leader and followers in the established master-follower game model. Their strategies are as follows; The strategy of a leading virtual power plant operator is a set of electricity and heat purchase and sale prices, which can be described as:

[0118] in , These represent the purchase and sale prices of electricity by the virtual power plant operator during time period t. , These represent the purchase and sale prices of heat for the virtual power plant operator during time period t. The strategy of a follower-based energy supply and demand coordination system is a set consisting of the output of each unit on the supply side, the transferable electrical load response, the reduceable thermal load response, and the energy storage output plan on the user side, which can be expressed in the following form:

[0119] When virtual power plant operators and energy supply and demand coordination systems reach equilibrium, it is denoted as... Under this equilibrium, neither party will benefit from unilaterally changing its strategy, which can be expressed as:

[0120] Step S8: Design a hierarchical solution strategy for the established two-layer optimization model. An improved genetic algorithm is used to solve the upper-layer VPPO model; while a quadratic programming, entropy weight method, and particle swarm optimization algorithm are used to jointly solve the lower-layer ESDS multi-objective optimization model. The upper and lower layers interact through their respective strategies.

[0121] Step S9: Input the preprocessed electricity and heat load datasets from the demand side and the typical scenario wind and solar power output data generated in Step S2 into the virtual power plant two-layer optimization model established in Step S6. Use the aforementioned layered solution strategy to obtain the optimized scheduling result. The optimized scheduling scheme obtained from solving the model includes the electricity and heat output power of the energy supply side in the ESDS after optimized scheduling in time period t, and the electricity and heat power of the user-side loads participating in demand response in time period t.

[0122] The specific demand-side preprocessed electricity and heat load data and the typical scenario wind and solar power output data generated in step S2 are as follows: Figure 3 As shown. The hierarchical solution strategy described in step S8 is as follows: Figure 4 As shown.

[0123] By inputting electricity and heat load data and wind and solar power output data from typical scenarios, specific optimized scheduling results are obtained, such as... Figure 5 , 6 , and as shown in Table 1.

[0124] Table 1 Revenue and Costs of VPPO

[0125] Table 1 shows the revenue and costs of each component under the scenario where Virtual Power Plant Operator (VPPO) maximizes its profits. As can be seen from the table, the revenue from selling energy by a VPPO is approximately twice the cost of purchasing energy. The difference between the cost of purchasing energy from the grid and the revenue from selling surplus electricity is RMB 1570.6.

[0126] The specific optimized scheduling results are as follows: Figure 5 , 6 As shown. Considering the low-carbon and economical nature of new energy sources, and the introduction of energy storage on the demand side in the Energy Supply and Demand Coordination System (ESDS), wind and solar power output on the supply side will be preferentially sold to VPPO. From 0:00 to 7:00, the overall output of wind and solar power is insufficient to meet the electricity demand on the demand side. Furthermore, due to the electricity consumption of electric boilers for heating during this period, gas turbines serve as backup power to compensate for some of the power shortage. From 11:00 to 16:00, due to high system electricity demand and low heat demand, photovoltaic and wind power show a complementary trend, with their combined output exceeding 450 kW and gas turbine power generation exceeding 300 kW. Since electric boilers are not operating, the available electricity on the supply side of the ESDS is between 800-1050 kW. The available heat is provided through the recovery of waste heat generated by gas turbines and heating by gas boilers. The peak electricity consumption period is from 17:00 to 22:00. Photovoltaic power output stops after 19:00, while wind power output increases significantly, with an average output of 531.8KW. Gas turbine power output averages 385.2KW, and the average saleable electricity is 917KW.

[0127] During the periods of 0:00-10:00, 16:00-17:00, and 22:00-23:00, due to the electrothermal coupling relationship and the fact that the combined heat energy from the gas-fired boiler's heating and recovery cannot meet the demand, the operation of the electric boiler is adjusted to supplement some of the heat. Throughout the entire scheduling cycle, the output of the gas-fired boiler ranges from 241-355 kW. This is because the heating cost of the gas-fired boiler is a quadratic function; the heating cost for the portion of the gas-fired boiler's output exceeding the current planned output is higher than the additional cost of electric heating.

[0128] Under the optimal pricing strategy of the leader VPPO, the weights of economic and low-carbon objectives in the follower's overall objectives are 0.62 and 0.38, respectively. This shows that under the current pricing strategy, ESDS places more emphasis on economic benefits, followed by low-carbon and environmental protection.

[0129] Specific VPPO pricing strategies include: Figure 7 , 8 As shown. Figure 7 The changes in electricity prices reveal the dynamic adjustments in electricity supply and demand throughout the day. During peak consumption periods (11:00-14:00 and 18:00-22:00), the VPPO (Power Utilization Policy Office) increases electricity sales prices to curb demand-side electricity consumption during peak hours; during off-peak periods (23:00-9:00), it lowers electricity sales prices to encourage increased demand-side electricity consumption. This helps balance grid load, improves electricity utilization during off-peak hours, and reduces energy waste. The VPPO incentivizes producers and consumers to adjust their electricity consumption behavior rationally through pricing mechanisms, achieving peak shaving and valley filling, and improving the economy and efficiency of the entire power system.

[0130] Figure 8 The changes in various heat prices reveal that heat prices are relatively stable during peak heat load periods, while fluctuating more significantly during off-peak periods. The relationship between the VPPO's purchase and sale heat prices and heat load is achieved through flexible electric boiler regulation, heat storage and release strategies, and load reduction responses to balance supply and demand and optimize costs. During peak heat load periods, the VPPO avoids price fluctuations by meeting demand; during off-peak heat load periods, it incentivizes demand by adjusting heat prices and utilizes off-peak electricity prices for heat storage to improve the system's energy efficiency and economics.

[0131] Example 2: This invention provides a computer device comprising a processor, a memory, a network interface, and other core hardware components. The memory stores the instruction set, the network interface enables data interaction with various devices within the virtual power plant, and the processor executes the instructions stored in the memory, thereby driving the device to run the aforementioned master-slave game-theoretic multi-objective optimization scheduling method for virtual power plants.

[0132] The preferred embodiments of the present invention have been described in detail above. The specific implementation examples presented are merely one example among many feasible solutions. Those skilled in the art are encouraged to build upon this, combining it with existing technologies, and through rigorous logical analysis, scientific reasoning, or reasonable and limited experimental exploration, develop more technical solutions adapted to different scenarios. These derivative solutions, although potentially differing in details, are all covered by the scope of protection defined by the claims of this invention, as long as their core ideas originate from this invention.

[0133] Example 3: This invention also provides a virtual power plant two-layer optimization scheduling device driven by master-slave game theory and multi-objective fusion, corresponding to Embodiment 1. At the hardware level, this black-start partitioning device for distribution networks based on multiple types of distributed power sources includes a processor, internal bus, network interface, memory, and non-volatile memory, and may also include other hardware required for other operations. The processor reads the corresponding computer program from the non-volatile memory into memory and then runs it to achieve the above-mentioned... Figure 1 The data acquisition method described above. Of course, in addition to software implementation, this invention does not exclude other implementation methods, such as logic devices or a combination of hardware and software, etc. That is to say, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0134] Improvements in a technology can be clearly distinguished as either hardware improvements (e.g., improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the methodology). However, with technological advancements, many improvements to the methodology can now be considered direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved methodology into the hardware circuit. Therefore, it cannot be said that an improvement in methodology cannot be implemented using hardware physical modules. For example, a Programmable Logic Device (PLD) (such as a Field Programmable Gate Array (FPGA)) is such an integrated circuit whose logic function is determined by the user programming the device. Designers can program and "integrate" a digital system onto a PLD themselves, without needing chip manufacturers to design and manufacture dedicated integrated circuit chips. Furthermore, nowadays, instead of manually manufacturing integrated circuit chips, this programming is mostly implemented using "logic compiler" software. Similar to the software compiler used in program development, the original code before compilation must also be written in a specific programming language, called a Hardware Description Language (HDL). There are many HDLs, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, and RHDL (Ruby Hardware Description Language). Currently, the most commonly used are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also understand that by simply performing some logic programming on the method flow using one of these hardware description languages ​​and programming it into an integrated circuit, the hardware circuit implementing the logical method flow can be easily obtained.

[0135] The controller can be implemented in any suitable manner. For example, it can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicon Labs C8051F320. A memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art will also recognize that, in addition to implementing the controller in purely computer-readable program code form, the same functionality can be achieved by logically programming the method steps to make the controller take the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the means included therein for implementing various functions can also be considered as structures within the hardware component. Alternatively, the means for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0136] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0137] Example 4: This invention also proposes a computer-readable storage medium on which a program is stored, which, when executed, implements the method described in Embodiment 1. Computer-readable media include permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient media, such as modulated data signals and carrier waves.

[0138] The various embodiments in this invention are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

Claims

1. A two-layer optimal scheduling method for virtual power plants driven by master-slave game theory and multi-objective fusion, characterized in that the method... include: Acquire preprocessed electricity and heat load datasets, as well as wind and solar power output data for typical scenarios; A two-layer multi-objective optimization framework is constructed, incorporating a master-slave game between a Virtual Power Plant Operator (VPPO) and an Energy Supply and Demand Coordination System (ESDS). The upper layer of the framework is the VPPO model, and the lower layer is the ESDS multi-objective optimization model. The ESDS multi-objective optimization model includes economic and environmental objectives and their corresponding weights. The VPPO model includes the electrical and thermal output power of the energy supply side in the ESDS after optimized scheduling in time period t, and the electrical and thermal power of the user-side loads participating in demand response in time period t. A hierarchical solution strategy is adopted to solve the two-layer optimization framework and obtain the optimized scheduling result.

2. The virtual power plant two-layer optimization scheduling method driven by master-slave game theory and multi-objective fusion as described in claim 1, characterized in that, The Virtual Power Plant Operator (VPPO) model is as follows: ； ； in, Let T represent the objective function of the Virtual Power Plant Operator (VPPO) model, where T is the number of time periods in a scheduling cycle. This represents the revenue that the virtual power plant operator receives from selling energy to the user side during time period t; This represents the energy supply revenue of the energy supply and demand coordination system during time period t; This represents the cost of VPPO purchasing energy from the energy supply side; This represents the amount of electricity traded between VPPO and the power grid. This indicates the heat load interruption penalty cost incurred by the VPPO due to insufficient heating supply to the user side. This represents the user's energy purchase cost. This represents the energy supply revenue of ESDS during time period t. , These represent the electrical and thermal power of the user-side load after participating in demand response during time period t; , These represent the electrical and thermal output power on the power supply side of the ESDS after optimized scheduling in time period t; For the transaction volume between VPPO and the power grid, The heat load interruption supply during period t; This is the penalty coefficient for heating interruption; , These are the purchase and sale prices of VPPO for time period t. , These are the purchase and sale prices of VPPO for time period t, respectively. , These represent the time-of-use electricity price and the grid connection price for the power grid during time period t, respectively.

3. The virtual power plant two-layer optimization scheduling method driven by master-slave game theory and multi-objective fusion as described in claim 2, characterized in that, After time period t optimization scheduling, the electrical and thermal output power on the power supply side of ESDS is: ； in, , These represent the electrical and thermal output power on the power supply side of the ESDS after optimized scheduling in time period t. For the efficiency of the heat exchanger, Provide photovoltaic power output for period t; Contribute to the scenery during period t; For the gas turbine output during time period t; The power consumption of the electric boiler during time period t. The heat power generated by the waste heat boiler during time period t; The output thermal power of the gas-fired boiler during time period t; The output heat power of the electric boiler during time period t.

4. The virtual power plant two-layer optimization scheduling method driven by master-slave game theory and multi-objective fusion as described in claim 3, characterized in that, The electrical and thermal power of the user-side load after participating in demand response during time period t are: ； in, , These represent the electrical and thermal power of the user-side load after participating in demand response during time period t. The original electrical load during time period t; Let t represent the response quantity of transferable electrical loads participating in demand response during time period t. , The charging and discharging power of the electrical energy storage during time period t. The original heat load for time period t; The response amount of the heat load that can be reduced during time period t; The heat storage and release power during time period t is the heat storage power.

5. The virtual power plant two-layer optimization scheduling method driven by master-slave game theory and multi-objective fusion as described in claim 4, characterized in that, The multi-objective optimization model for the Energy Supply and Demand Coordination System (ESDS) is as follows: ； in, , The weights for economic and environmental objectives are respectively. and These are economic objectives and environmental objectives.

6. The virtual power plant two-layer optimization scheduling method driven by master-slave game theory and multi-objective fusion as described in claim 5, characterized in that, The economic objective is: ； in, This represents the utility function of users during time period t, i.e., user energy satisfaction. The total fuel cost on the energy supply side, This represents the total fuel cost on the user's side.

7. The virtual power plant two-layer optimization scheduling method driven by master-slave game theory and multi-objective fusion as described in claim 6, characterized in that, Environmental objectives are: ； in, Cost of carbon dioxide treatment; , These are the carbon emission coefficients for gas turbines and gas boilers, respectively. Let t represent the utility function of users during time period t, i.e., user energy satisfaction.

8. The virtual power plant two-layer optimization scheduling method driven by master-slave game theory and multi-objective fusion as described in claim 7, characterized in that, User energy consumption satisfaction is: ； in, , This represents the user's preference coefficient for energy consumption. , This represents the user's preference coefficient for consuming heat energy.

9. The virtual power plant two-layer optimization scheduling method driven by master-slave game theory and multi-objective fusion as described in claim 1, characterized in that, The preprocessing includes outlier detection, correction, and missing value compensation.

10. The virtual power plant two-layer optimization scheduling method driven by master-slave game theory and multi-objective fusion as described in claim 9, characterized in that, The specific steps for missing value compensation are to fill in the missing values ​​using linear interpolation.

Images