Power system multi-agent collaborative optimization scheduling method considering virtual energy storage

Abstract

The invention relates to a power system multi-agent collaborative optimization scheduling method considering virtual energy storage, and the method comprises the steps: building a double-layer optimization model which takes a power distribution network operation unit as a leader, and takes an energy storage operator and a park integrated energy system as followers, optimizing the day-ahead electricity price, and reducing the cost of the power distribution network operation unit. A day-ahead-intra-day two-stage optimization scheduling model taking a power distribution network operation unit, an energy storage operator and a park integrated energy system as a main body is constructed, and a scheduling strategy of charging and discharging of the energy storage operator is comprehensively optimized in a day-ahead stage; and dynamically correcting the wind and light output and load deviation in the intraday stage. And a day-ahead-intra-day two-stage optimization scheduling model is adopted, so that the operation cost and the carbon emission intensity of the power system are reduced. By constructing a dynamic virtual energy storage model of an electric vehicle cluster, the'load-storage 'flexible adjustment capability of the electric vehicle cluster is quantified, the dependence on a peak regulation unit is reduced, and the flexibility and economy of a power system are comprehensively improved.

Description

Technical Field

[0001] This invention belongs to the field of power system optimization and dispatching technology, specifically relating to a multi-entity collaborative optimization and dispatching method for power systems that considers virtual energy storage. Background Technology

[0002] Park-level integrated energy system (PIES) is an important practical direction for the transformation of energy systems towards low-carbon and intelligent transformation, and can integrate multiple energy forms to achieve multi-energy synergistic optimization.

[0003] Electric vehicles (EVs), as energy-consuming devices with dual "load-storage" properties, can be clustered in large numbers to act as virtual energy storage (VES) resources and form a multi-energy complementary and synergistic mechanism with other energy storage resources in the integrated energy system, thereby optimizing energy allocation at the spatiotemporal levels.

[0004] The paper "Multi-microgrid Hybrid Game Operation Strategy Considering Virtual Energy Storage Participation" published in the fourth issue of "Power System Technology" in 2025 by Wu Ruixing et al. discloses an optimal scheduling model for integrated energy systems including electric vehicles. It adopts a two-level optimization method that considers energy storage charging and discharging strategies and system costs to improve the system's economy and stability.

[0005] The paper "Optimization of Multi-dimensional Peak Shaving Auxiliary Services Supported by Virtual Energy Storage of Electric Vehicles" published in the S1 issue of the Proceedings of the Chinese Society for Electrical Engineering in 2024, by Hou Hui et al., builds on the above and constructs a compensation mechanism that considers the peak shaving contribution of electric vehicles, so as to treat electric vehicles as a flexible resource to participate in peak shaving, thereby effectively reducing the peak-valley difference and peak shaving cost of the system.

[0006] The paper "Research on Low-Carbon Operation Optimization of Integrated Energy System in Industrial Park Considering Hybrid Virtual Energy Storage," published in the second issue of *Thermal Power Generation* in 2025 by Zhao Zhenyu et al., discloses an integrated energy system incorporating dynamic virtual energy storage incentive mechanisms and carbon cycle mechanisms. It utilizes hybrid energy storage to optimize system operating costs and reduce carbon emissions. While this literature optimizes the scheduling of integrated energy systems containing VES (Virtual Energy Storage), it neglects the interests of VES as an independent entity.

[0007] The aforementioned literature has conducted some research on the optimal scheduling methods for integrated energy systems including VES, but it still has the following shortcomings. On the one hand, focusing only on a single-agent optimization framework makes it difficult to coordinate the benefit balance problem when multiple agents participate; on the other hand, for multi-agent collaborative scheduling problems, existing methods fail to consider the impact of day-ahead electricity price changes on intraday optimization results under source-load cooperation.

[0008] The difficulty in solving the above technical problems: When Virtual Energy Storage (VES) operates as an independent entity, its interests must align with the overall optimization goals of the Park Integrated Energy System (PIES). Traditional models treat VES as a passive resource, neglecting its autonomous decision-making power as a market participant. The coupling uncertainty between wind and solar power output and load demand dynamically evolves in both the day-ahead and intraday phases, necessitating the construction of a cross-timescale collaborative scheduling framework. Existing research often focuses on a single entity or a single timescale, making it difficult to achieve dynamic responses from multiple entities at different stages. Summary of the Invention

[0009] The purpose of this invention is to address the aforementioned problems by providing a multi-entity collaborative optimization scheduling method for power systems that considers virtual energy storage. This method constructs a two-stage optimization scheduling model with distribution network operators, energy storage operators, and integrated energy systems in industrial parks as the main entities. In the day-ahead stage, it comprehensively optimizes the charging and discharging scheduling strategies of energy storage operators. In the intraday stage, it dynamically corrects the deviation between wind and solar power output and load through rolling optimization, thereby reducing the operating costs of the distribution network system.

[0010] To achieve the above objectives, the technical solution provided by this invention is as follows: A multi-entity collaborative optimization scheduling method for power systems considering virtual energy storage is proposed. A two-layer optimization model is constructed with the distribution network operator as the leader and the energy storage operator and the park's integrated energy system as followers to optimize day-ahead electricity prices and reduce the costs of the distribution network operator.

[0011] The method includes the following steps: Step 1: Establish a virtual energy storage model for electric vehicle clusters. For the step power generated by multiple electric vehicles being connected to or disconnected from the grid at the same time, introduce deviation power to correct the virtual energy storage model. Step 2: Construct a two-stage collaborative optimization model that includes a day-ahead optimization model and an intraday optimization model. The optimization objectives of the current phase optimization model include the energy consumption cost of the park's integrated energy system and the operating cost of the energy storage operator; The optimization objectives of the intraday phase optimization model include the marginal benefits of the distribution network operator and the overall electricity purchase cost of the energy storage operator and the park's integrated energy system; Step 3: Solve the two-stage collaborative optimization model to obtain the optimal solution, i.e., the optimal scheduling scheme; Step 4: Based on the optimal scheduling scheme obtained in Step 3, guide the optimized operation of the power system.

[0012] Furthermore, in step 1, it is assumed that the grid connection time, grid disconnection time, and initial energy of electric vehicles in the park's integrated energy system all follow a normal distribution, and that the energy level of electric vehicles when they disconnect from the grid is above the expected energy level. Considering the uncertainty of the grid connection and grid disconnection time of electric vehicles, a state parameter is introduced. Cluster control of electric vehicles , The mathematical model for a single electric vehicle is as follows: (Descriptions of the grid-connected and off-grid states of the electric vehicle are provided.) (1) In the formula: , , , , , These are the battery capacity, power interaction with the grid, charging power, discharging power, maximum battery capacity, and minimum battery capacity of the i-th electric vehicle, respectively. , These are the charging and discharging efficiencies of electric vehicles, respectively. , These represent the maximum charging and discharging power of the electric vehicle, respectively. , For the grid connection and off-grid times of electric vehicles; , These represent the off-grid power consumption of electric vehicles and the minimum expected power consumption, respectively. To adjust the step size; The schedulable power boundary and battery capacity boundary of the electric vehicle cluster are: (2) (3) In the formula: , , , Let represent the maximum charging power, maximum discharging power, maximum battery capacity, and minimum battery capacity of the electric vehicle cluster at time t, respectively; N is the total number of electric vehicles.

[0013] Preferably, in step 1, in order to address the step power fluctuations caused by multiple electric vehicles simultaneously connecting to or disconnecting from the grid, a deviation power is introduced to correct for the fluctuations. (4) In the formula: , These represent the battery levels of the i-th electric vehicle when it is connected to the grid and when it is disconnected from the grid, respectively.

[0014] The virtual energy storage model for electric vehicle clusters is as follows: (5) In the formula: , , , These are the energy storage capacity of the electric vehicle cluster, the power exchange with the power grid, the charging power, and the discharging power.

[0015] In step 2, to address the impact of day-ahead source-load uncertainty, a two-stage optimization scheduling model considering the benefits of each entity is constructed based on the carbon emission cost mechanism. The objective function of the day-ahead optimization model includes minimizing the overall cost for the energy storage operator. (6) (7) In the formula: Costs for energy storage operators; , These are the costs of physical energy storage and the costs of virtual energy storage, respectively. , , These are the system electricity price, electric vehicle charging service fee, and energy storage electricity sales price. The electricity sales price for the park's integrated energy system; , This refers to the power volume that the energy storage operator purchases from the distribution network operator and the power volume that it sells to the park's integrated energy system; The electricity sold by the park's integrated energy system to energy storage operators; This refers to the energy storage charging and discharging cost coefficient. This is the discharge cost coefficient for electric vehicles.

[0016] Furthermore, the objective function of the current-stage optimization model also includes minimizing the energy cost of the park's integrated energy system. The park's integrated energy system includes photovoltaic generators, wind turbines, combined cooling, heating and power (CCHP) units, power-to-gas and carbon capture coupling devices, gas boilers, electric boilers, and electric chillers, and also incorporates energy storage equipment; through multi-energy complementarity, it can fully balance energy supply and demand, ensuring the flexibility of the park's integrated energy system.

[0017] (8) (9) In the formula: , , , These are the gas purchase cost, carbon emission cost, electricity purchase cost, and operation and maintenance cost of the park's integrated energy system. Electricity is purchased from the power distribution network operator for the park's integrated energy system; , These are the system's time-of-use gas price and gas purchase capacity, respectively. , Energy storage devices The charging and discharging power This is the charging and discharging cost coefficient; , The operation and maintenance cost coefficient and output power of energy conversion device x.

[0018] Preferably, in step 2, a carbon cost model is established within the integrated energy system, taking into account the low-carbon nature of the park's integrated energy system; to reduce system carbon emissions, a tiered carbon emission cost mechanism is adopted to calculate the carbon emission cost price based on the actual net carbon emissions, and the calculation of carbon emissions and carbon emission cost price is improved. (10) In the formula: , These are the carbon emission costs and actual net emissions, respectively. , These are the base carbon price and the carbon price growth rate, respectively. The length of the carbon content increase interval; (11) (12) In the formula: , These are actual carbon emissions and carbon emission quotas, respectively. , These are the actual carbon emission coefficient and carbon emission quota per unit of electricity; , , , The actual carbon emission coefficient and carbon emission quota for gas-fired boilers and combined cooling, heating and power units; This represents the carbon absorption coefficient during gas production in the methane reactor. , These are the conversion coefficients for electricity to heat and cold to heat, respectively; , These are the heat output power and gas power of the gas-fired boiler, respectively. , , These refer to the power output of the combined cooling, heating and power (CCHP) unit, specifically the power generated by electricity, heat, and cooling. Gas purchase capacity for combined cooling, heating and power units; This is the power consumption of the methane reactor.

[0019] In step 2, the objective function of the upper layer of the two-level optimization model is to minimize the unit cost of distribution network operation. (13) In the formula: , These are the electricity sales revenue and electricity purchase cost for the power distribution network operator. Price in response to the needs of the park; (14) In the formula: The price at which distribution network operators purchase electricity from the superior power grid; The optimized system electricity price; This refers to the power consumption of the park after demand response; , These are the compensation coefficients for energy that can be reduced and energy that can be transferred, respectively. , These are the power that can be reduced and the power that can be transferred, respectively.

[0020] The constraints of the upper layer of the two-layer optimization model are: (15) In the formula: , , These represent the minimum, maximum, and average electricity sales prices of power distribution network operators, respectively.

[0021] The lower layer of the two-layer optimization model aims to minimize the overall electricity purchase cost, simultaneously optimizing the revenue of the energy storage operator and the overall energy cost of the park's integrated energy system based on electricity prices and incentive signals. (16) (17) In the formula, , These are the electricity purchase costs from the power distribution network operator and the integrated energy system of the park, respectively.

[0022] The constraints of the lower level of the two-level optimization model include: Demand response constraints: (18) Power balance constraints: (19) In step 2, the intraday phase collaborative optimization model performs rolling optimization with the objective of minimizing the comprehensive cost of the energy storage operator and the park's integrated energy system within the current rolling time domain, resulting in an intraday optimized scheduling scheme. Its objective function is: (20) ;(twenty one) In the formula: , These are the costs of physical energy storage and virtual energy storage within the day, respectively. , , , These are the gas purchase cost, carbon emission cost, electricity purchase cost, and operation and maintenance cost of the park's integrated energy system within the day; , These are the intraday charging price for electric vehicles and the electricity sales price for energy storage. The electricity sales price for the park's integrated energy system; This refers to the daily gas purchase volume of the system. , These are the charging and discharging power of the intraday thermal storage device, , These are the charging and discharging power of the cold storage device, respectively. Energy conversion equipment Daily output power.

[0023] Compared with the prior art, the beneficial effects of the present invention include: (1) By constructing a two-layer optimization model with the distribution network operator as the upper layer and the energy storage operator and the park integrated energy system (PIES) as the lower layer, and designing a dynamic electricity price signal transmission mechanism, this invention achieves dynamic equilibrium of benefits for multiple parties. The distribution network operator guides the energy storage operator and the park integrated energy system PIES to optimize charging and discharging strategies by adjusting the electricity price, while ensuring the benefits of VES as an independent entity. This solves the pain points of "uneven distribution of benefits" and "conflict of demands between entities" in the traditional model and improves the overall system economy.

[0024] (2) A hierarchical decision-making framework of "day-ahead decision-intraday correction" is adopted. The energy storage dispatch strategy is pre-optimized in the day-ahead stage. In the intraday stage, the deviation between wind and solar power output and load is dynamically corrected through rolling optimization to reduce the impact of prediction errors on the operating results. This model further couples carbon emission costs, thereby reducing the operating costs and carbon emission intensity of the power system.

[0025] (3) By constructing a dynamic virtual energy storage model with electric vehicle clusters as the core, its "load-storage" flexible adjustment capability is quantified and complemented by physical energy storage, CCHP and other multi-energy devices. Under the incentive of electricity prices, EV clusters can charge during off-peak hours and discharge during peak hours, reducing electricity costs, reducing DNO's dependence on peak-shaving units, and comprehensively improving system flexibility and economy. Attached Figure Description

[0026] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0027] Figure 1 This is a framework diagram of the two-layer optimization model provided in the embodiments of the present invention.

[0028] Figure 2 This is a schematic diagram of the two-stage collaborative optimization model provided in the embodiments of the present invention.

[0029] Figure 3 This is a schematic diagram of the power distribution network operator, energy storage operator, and integrated energy system of the park in this embodiment of the invention.

[0030] Figure 4 This is a bar chart showing the day-ahead electrical load optimization results provided in this embodiment of the invention.

[0031] Figure 5 This is a bar chart showing the day-ahead heat load optimization results provided in this embodiment of the invention.

[0032] Figure 6 This is a line graph of intraday electricity prices provided in an embodiment of the present invention.

[0033] Figure 7 This is a line graph of the demand response results provided in an embodiment of the present invention.

[0034] Figure 8 This is a graph showing the results of the daily electricity load optimization provided in this embodiment of the invention.

[0035] Figure 9 This is a graph showing the results of the daily heat load optimization provided in this embodiment of the invention. Detailed Implementation

[0036] like Figure 1 As shown, a multi-agent collaborative optimization scheduling method for power systems considering virtual energy storage includes: Step 1: Establish a virtual energy storage model for electric vehicle clusters. For the step power generated by multiple electric vehicles being connected to or disconnected from the grid at the same time, introduce deviation power to correct the virtual energy storage model. Assume there are 200 EVs in the PIES, and that the EVs' grid entry time, grid exit time, and initial battery level all follow a normal distribution, ensuring that the battery level of an EV is above the expected level when it leaves the grid. Considering the uncertainty of EV grid entry and exit times, a state parameter is introduced. Cluster control of EVs The values ​​1 and 0 respectively describe the grid-connected and off-grid states of the EV.

[0037] The model for individual EVs is as follows: (1) In the formula: , , , , , These are the battery capacity, grid interaction power, charging power, discharging power, and maximum and minimum battery capacity of the i-th EV, respectively. , For the charging and discharging efficiency of EVs; , Indicates the maximum charging and discharging power of the EV; , For EV grid connection and off-grid times; , This indicates the EV's off-grid power consumption and minimum expected power consumption. To adjust the step size.

[0038] Since the grid connection and off-grid times of EVs within the park are relatively uniform, they can be aggregated as an EV cluster for comprehensive control. The schedulable power boundary and battery capacity boundary of the EV cluster are: (2) (3) In the formula: , , , These represent the maximum charging and discharging power and the maximum and minimum battery capacity of the EV cluster at time t, respectively; N is the total number of electric vehicles.

[0039] To address the step-level power surge caused by multiple electric vehicles simultaneously connecting to or disconnecting from the grid, a deviation power is introduced for correction, as shown in equation (4). (4) In the formula: , The battery level of the i-th EV when it is connected to or disconnected from the grid.

[0040] The virtual energy storage model of the EV cluster is shown in equation (5). (5) In the formula: , , , These are the energy storage capacity of the EV cluster, the power of interaction with the grid, the charging power, and the discharging power.

[0041] Step 2: Construct a two-stage collaborative optimization model that includes a day-ahead optimization model and an intraday optimization model. The optimization objectives of the day-ahead optimization model include the energy consumption cost of the park's integrated energy system and the operating cost of the energy storage operator. The optimization objectives of the intraday optimization model include the marginal benefits of the distribution network operator and the overall electricity purchase cost of the energy storage operator and the park's integrated energy system. To address the impact of day-ahead source-load uncertainty, a two-stage optimization scheduling model considering the benefits of each entity is constructed based on the carbon emission cost mechanism. The objective function of the day-ahead optimization model includes minimizing the overall cost for energy storage operators. (6) (7) In the formula: Costs for energy storage operators; , These are the costs of physical energy storage and the costs of virtual energy storage, respectively. , , These are the system electricity price, electric vehicle charging service fee, and energy storage electricity sales price. The electricity sales price for the park's integrated energy system; , This refers to the power volume that the energy storage operator purchases from the distribution network operator and the power volume that it sells to the park's integrated energy system; The electricity sold by the park's integrated energy system to energy storage operators; This refers to the energy storage charging and discharging cost coefficient. This is the discharge cost coefficient for electric vehicles.

[0042] In this embodiment, the objective function of the day-ahead optimization model also includes minimizing the energy cost of the park's integrated energy system. The park's integrated energy system includes photovoltaic generators, wind turbines, combined cooling, heating and power (CCHP) units, power-to-gas and carbon capture coupling devices, gas boilers, electric boilers, and electric chillers, and also includes energy storage equipment; through multi-energy complementarity, it can fully balance energy supply and demand, ensuring the energy flexibility of the park's integrated energy system.

[0043] (8) (9) In the formula: , , , These are the gas purchase cost, carbon emission cost, electricity purchase cost, and operation and maintenance cost of the park's integrated energy system. Electricity is purchased from the power distribution network operator for the park's integrated energy system; , These are the system's time-of-use gas price and gas purchase capacity, respectively. , Energy storage devices The charging and discharging power This is the charging and discharging cost coefficient; , The operation and maintenance cost coefficient and output power of energy conversion device x.

[0044] In this embodiment, a carbon cost model is established within the integrated energy system, taking into account the low-carbon nature of the park's integrated energy system. To reduce system carbon emissions, a tiered carbon emission cost mechanism is adopted to calculate the carbon emission cost price based on the actual net carbon emissions, and the calculation of carbon emissions and carbon emission cost prices is improved. (10) In the formula: , These are the carbon emission costs and actual net emissions, respectively. , These are the base carbon price and the carbon price growth rate, respectively. The length of the carbon content increase interval; (11) (12) In the formula: , These are actual carbon emissions and carbon emission quotas, respectively. , These are the actual carbon emission coefficient and carbon emission quota per unit of electricity; , , , The actual carbon emission coefficient and carbon emission quota for gas-fired boilers and combined cooling, heating and power units; This represents the carbon absorption coefficient during gas production in the methane reactor. , These are the conversion coefficients for electricity to heat and cold to heat, respectively; , These are the heat output power and gas power of the gas-fired boiler, respectively. , , These refer to the power output of the combined cooling, heating and power (CCHP) unit, specifically the power generated by electricity, heat, and cooling. Gas purchase capacity for combined cooling, heating and power units; This is the power consumption of the methane reactor.

[0045] In the embodiment, a two-layer optimization model is established with the distribution network operator as the leader and the energy storage operator and the park's integrated energy system as followers to optimize the day-ahead electricity price and reduce the cost of the distribution network operator. The objective function of the upper layer of the two-level optimization model is to minimize the unit cost of distribution network operation. (13) In the formula: , These are the electricity sales revenue and electricity purchase cost for the power distribution network operator. Price in response to the needs of the park; (14) In the formula: The price at which distribution network operators purchase electricity from the superior power grid; The optimized system electricity price; This refers to the power consumption of the park after demand response; , These are the compensation coefficients for energy that can be reduced and energy that can be transferred, respectively. , These refer to the power that can be reduced and the power that can be transferred, respectively. The constraints of the upper layer of the two-layer optimization model are: (15) In the formula: , , These represent the minimum, maximum, and average electricity sales prices of power distribution network operators, respectively. The lower layer of the two-layer optimization model aims to minimize the overall electricity purchase cost, simultaneously optimizing the revenue of the energy storage operator and the overall energy cost of the park's integrated energy system based on electricity prices and incentive signals. (16) (17) In the formula, , These are the electricity purchase costs from the power distribution network operator and the integrated energy system of the park, respectively. The constraints of the lower level of the two-level optimization model include: Demand response constraints: (18) Power balance constraints: ;(19)

[0046] In this embodiment, the intraday phase collaborative optimization model aims to minimize the overall cost of the energy storage operator and the park's integrated energy system within the current rolling time domain, thereby obtaining an intraday optimized scheduling scheme. Its objective function is: (20) ;(twenty one) In the formula: , These are the costs of physical energy storage and virtual energy storage within the day, respectively. , , , These are the gas purchase cost, carbon emission cost, electricity purchase cost, and operation and maintenance cost of the park's integrated energy system within the day; , These are the intraday charging price for electric vehicles and the electricity sales price for energy storage. The electricity sales price for the park's integrated energy system; This refers to the daily gas purchase volume of the system. , These are the charging and discharging power of the intraday thermal storage device, , These are the charging and discharging power of the cold storage device, respectively. Energy conversion equipment Daily output power.

[0047] Step 3: Solve the two-stage collaborative optimization model to obtain the optimal solution, i.e., the optimal scheduling scheme; Step 4: Based on the optimal scheduling scheme obtained in Step 3, guide the optimized operation of the power system.

[0048] Calculation example: This paper studies the multi-entity collaborative optimization strategy within a power distribution network system, constructing a two-stage collaborative optimization model involving power distribution network operators, energy storage operators, and a comprehensive energy system in a park. To verify the economic efficiency and low-carbon performance of the proposed model during system optimization, a simulation analysis is conducted using a comprehensive energy system in a park in Sichuan Province as an example. The Monte Carlo method is used to generate the electric vehicle charging demand within the park. The model parameters, time-of-use electricity pricing, and time-of-use gas pricing schemes proposed in this invention are all derived from relevant literature. It is assumed that the daily renewable energy output and load follow the probability distribution of the day-ahead data, and the data error follows a Gaussian noise distribution.

[0049] In terms of power balance, by Figure 4 It can be seen that the system exhibits differentiated energy supply characteristics at different times. During the off-peak hours at night, from 0:00 to 03:00, PIES electricity consumption is low, and wind power output can cover the system's energy demand, achieving zero external power purchase. From 04:00 to 08:00, PIES electricity consumption surges, and renewable energy output cannot meet the park's energy demand, so it is necessary to purchase electricity from the DNO to meet the park's energy demand. From 12:00 to 15:00, solar power resources are abundant and can meet PIES electricity demand. The system sells excess electricity to energy storage operators to reduce its own electricity costs and uses energy conversion devices to convert electricity into heat, cold, hydrogen, and natural gas. Heat and cold energy are stored in multi-energy storage devices for use when energy prices are high, while hydrogen and natural gas are used to reduce the system's gas purchase costs. From 16:00 to 24:00, affected by the decline in photovoltaic output, it is necessary to purchase electricity from the DNO and energy storage operators to make up for the insufficient PIES electricity.

[0050] In terms of thermal power balance, by Figure 5It can be seen that the heat load demand in the system is met through the coordinated regulation of EB, GB, CCHP units and energy storage system. During periods of high gas prices in the system, the thermal power output of GB decreases, and the heat released by EB and thermal storage meets the heat gap in PIES. From 10:00 to 12:00 and from 14:00 to 16:00, the output of renewable energy is relatively large. The system uses EB to convert electrical energy into thermal energy and store it in thermal storage devices, realizing the temporal and spatial transfer and utilization of energy, effectively mitigating the operating costs during subsequent periods of high energy prices.

[0051] like Figure 6-7 As shown, to improve DNO revenue, its pricing strategy is similar to the trend of power purchases by lower-level consortia. Between 00:00-10:00 and 18:00-22:00, the power purchases by lower-level consortia are relatively large; therefore, to improve DNO revenue, the system electricity price should be set at its upper limit. Between 10:00-14:00, renewable energy output can meet the system's energy demand; therefore, the power purchases by lower-level consortia are relatively small, and the electricity price will be reduced to lower the average electricity price.

[0052] Table 1. Intraday Optimization Results

[0053] In terms of power balance, by Figure 8 It can be seen that, considering the impact of intraday electricity price changes and demand response on PIES's electricity consumption plan, during the period from 00:00 to 08:00, the output of new energy sources within PIES cannot meet its own energy demand. At this time, the energy demand within the system relies on the joint supply from external DNO and energy storage operators. During the period from 10:00 to 15:00, the new energy sources within the system can cover its own energy demand, and the surplus electricity is converted into heat, cold, hydrogen, and gas resources for the park's use through energy conversion equipment. During the period from 20:00 to 24:00, since the electricity price of energy storage is lower than that of DNO, in order to reduce the park's energy costs, PIES mainly purchases electricity from energy storage operators to make up for the electricity gap.

[0054] In terms of economic benefits, Table 1 shows that PIES exhibits significant operational advantages in the two-stage multi-stakeholder collaborative optimization. During the intraday phase, PIES' energy purchase cost decreased by 7.92% compared to the daytime level, DNO revenue increased by 7.21%, and carbon emissions in the system decreased by 6329.96 kg, meeting both overall benefit requirements and environmental requirements. Due to the adjustment of the overall intraday electricity price, PIES optimized its energy consumption structure, resulting in increased overall electricity consumption and decreased gas consumption during the intraday phase. This helped reduce system carbon emissions while simultaneously increasing DNO operating revenue. However, the decrease in the intraday average electricity price narrowed the arbitrage space for energy storage operators' "low-storage, high-release" strategy, preventing them from earning high price differences and reducing their revenue by 6.89%. These results indicate that multi-stakeholder collaborative optimization requires balancing the conflicting relationship between local interests and overall optimization.

[0055] To address the optimization scheduling problem of integrated energy systems with multiple stakeholders in industrial parks under scenarios with high penetration of new energy sources, this invention proposes a two-stage collaborative optimization scheduling model for integrated energy systems in industrial parks, considering virtual energy storage and carbon emission cost mechanisms. The main conclusions are as follows: 1) The proposed two-stage scenario, which includes a three-entity collaborative optimization model involving PIES, DNO, and energy storage operators, can provide a reference for multi-entity collaborative optimization schemes. Using the method of this invention, PIES can reduce its energy costs by 7.67%, DNO revenue can be increased by 7.21%, and carbon emissions can be reduced by 17.08%, thus achieving low-carbon operation of the system.

[0056] 2) This invention introduces electric vehicles as VES to establish a two-way "load-storage" conversion mechanism and participate in system optimization scheduling. Results show that VES can reduce carbon emissions from the park's integrated energy system and increase the revenue of energy storage operators and DNOs by 46.47% and 25.13%, respectively.

[0057] 3) Adjusting electricity prices using optimization models in the two-stage optimization process can effectively stimulate the electricity consumption potential of PIES, assist in optimizing the system's energy consumption structure, reduce the electricity cost of PIES, and achieve the effects of increasing the revenue of energy storage operators and promoting the consumption of new energy.

Claims

1. A multi-agent collaborative optimization scheduling method for power systems considering virtual energy storage, characterized in that, By constructing a two-stage optimization scheduling model with distribution network operators, energy storage operators, and park integrated energy systems as the main entities, the charging and discharging scheduling strategies of energy storage operators are comprehensively optimized in the day-ahead stage. Dynamically adjust the deviation between wind and solar power output and load during the day to reduce the operating cost of the distribution network; The scheduling method includes the following steps: Step 1: Establish a virtual energy storage model for electric vehicle clusters. For the step power generated by multiple electric vehicles being connected to or disconnected from the grid at the same time, introduce deviation power to correct the virtual energy storage model. Step 2: Construct a two-stage collaborative optimization model that includes a day-ahead optimization model and an intraday optimization model. The optimization objectives of the current phase optimization model include the energy consumption cost of the park's integrated energy system and the operating cost of the energy storage operator; The optimization objectives of the intraday phase optimization model include the marginal benefits of the distribution network operator and the overall electricity purchase cost of the energy storage operator and the park's integrated energy system; Step 3: Solve the two-stage collaborative optimization model to obtain the optimal solution, i.e., the optimal scheduling scheme; Step 4: Based on the optimal scheduling scheme obtained in Step 3, guide the optimized operation of the power system.

2. The multi-entity collaborative optimization scheduling method for power systems according to claim 1, characterized in that, The aforementioned scheduling method establishes a two-layer optimization model with the distribution network operator as the leader and the energy storage operator and the park's integrated energy system as followers, optimizing day-ahead electricity prices and reducing the costs of the distribution network operator.

3. The multi-entity collaborative optimization scheduling method for power systems according to claim 2, characterized in that, In step 1, a virtual energy storage model for an electric vehicle cluster is constructed to promote the utilization of virtual energy storage resources. Assuming that the grid connection time, grid disconnection time, and initial energy of electric vehicles in the park's integrated energy system all follow a normal distribution, and guaranteeing that the energy level of electric vehicles upon grid disconnection is above the expected level, considering the uncertainty of the grid connection and grid disconnection times, a state parameter is introduced. Cluster control of electric vehicles , The mathematical model for a single electric vehicle is as follows: (Descriptions of the grid-connected and off-grid states of the electric vehicle are provided.) ； (1) In the formula: , , , , , These are the battery capacity, power interaction with the grid, charging power, discharging power, maximum battery capacity, and minimum battery capacity of the i-th electric vehicle, respectively. , These are the charging and discharging efficiencies of electric vehicles, respectively. , These represent the maximum charging and discharging power of the electric vehicle, respectively. , For the grid connection and off-grid times of electric vehicles; , These represent the off-grid power consumption of electric vehicles and the minimum expected power consumption, respectively. To adjust the step size; The schedulable power boundary and battery capacity boundary of the electric vehicle cluster are: ；(2) ；(3) In the formula: , , , Let represent the maximum charging power, maximum discharging power, maximum battery capacity, and minimum battery capacity of the electric vehicle cluster at time t, respectively; N is the total number of electric vehicles.

4. The multi-entity collaborative optimization scheduling method for power systems according to claim 3, characterized in that, In step 1, to address the issue of fluctuating battery power caused by multiple electric vehicles simultaneously connecting to or disconnecting from the grid, a deviation battery power is introduced for correction. ； (4) In the formula: , These represent the battery levels of the i-th electric vehicle when it is connected to the grid and when it is disconnected from the grid, respectively.

5. The multi-entity collaborative optimization scheduling method for power systems according to claim 4, characterized in that, The virtual energy storage model for electric vehicle clusters is as follows: ； (5) In the formula: , , , These are the energy storage capacity of the electric vehicle cluster, the power exchange with the power grid, the charging power, and the discharging power.

6. The multi-entity collaborative optimization scheduling method for power systems according to claim 5, characterized in that, In step 2, to address the impact of day-ahead source-load uncertainty, a two-stage optimization scheduling model considering the benefits of each entity is constructed based on the carbon emission cost mechanism. The objective function of the day-ahead optimization model includes minimizing the overall cost for the energy storage operator. ；（6） ；（7） In the formula: Costs for energy storage operators; , These are the costs of physical energy storage and the costs of virtual energy storage, respectively. , , These are the system electricity price, electric vehicle charging service fee, and energy storage electricity sales price. The electricity sales price for the park's integrated energy system; , This refers to the power volume that the energy storage operator purchases from the distribution network operator and the power volume that it sells to the park's integrated energy system; The electricity sold by the park's integrated energy system to energy storage operators; This refers to the energy storage charging and discharging cost coefficient. This is the discharge cost coefficient for electric vehicles.

7. The multi-entity collaborative optimization scheduling method for power systems according to claim 6, characterized in that, The objective function of the current stage optimization model also includes minimizing the energy cost of the park's integrated energy system; The park's integrated energy system includes photovoltaic generators, wind turbines, combined cooling, heating and power (CCHP) units, power-to-gas and carbon capture coupling devices, gas boilers, electric boilers, and electric chillers, and also includes energy storage equipment. Through multi-energy complementarity, it can fully balance energy supply and demand, ensuring the flexibility of energy use in the park's integrated energy system. ；（8） ； （9） In the formula: , , , These are the gas purchase cost, carbon emission cost, electricity purchase cost, and operation and maintenance cost of the park's integrated energy system. Electricity is purchased from the power distribution network operator for the park's integrated energy system; , These are the system's time-of-use gas price and gas purchase capacity, respectively. , Energy storage devices The charging and discharging power This is the charging and discharging cost coefficient; , The operation and maintenance cost coefficient and output power of energy conversion device x.

8. The multi-entity collaborative optimization scheduling method for power systems according to claim 7, characterized in that, In step 2, a carbon cost model for the integrated energy system is established, taking into account the low-carbon nature of the park's integrated energy system; To reduce system carbon emissions, a tiered carbon emission cost mechanism is adopted to calculate the carbon emission cost price based on the actual net carbon emissions, and the calculation of carbon emission volume and carbon emission cost price is improved. ；（10） In the formula: , These are the carbon emission costs and actual net emissions, respectively. , These are the base carbon price and the carbon price growth rate, respectively. The length of the carbon content increase interval; ；（11） ；（12） In the formula: , These are actual carbon emissions and carbon emission quotas, respectively. , These are the actual carbon emission coefficient and carbon emission quota per unit of electricity; , , , The actual carbon emission coefficient and carbon emission quota for gas-fired boilers and combined cooling, heating and power units; This represents the carbon absorption coefficient during gas production in the methane reactor. , These are the conversion coefficients for electricity to heat and cold to heat, respectively; , These are the heat output power and gas power of the gas-fired boiler, respectively. , , These refer to the power output of the combined cooling, heating and power (CCHP) unit, specifically the power generated by electricity, heat, and cooling. Gas purchase capacity for combined cooling, heating and power units; This is the power consumption of the methane reactor.

9. The multi-entity collaborative optimization scheduling method for power systems according to claim 8, characterized in that, In step 2, a two-layer optimization model is established with the distribution network operator as the leader and the energy storage operator and the park's integrated energy system as followers to optimize the day-ahead electricity price and reduce the cost of the distribution network operator. The objective function of the upper layer of the two-level optimization model is to minimize the unit cost of distribution network operation. ； （13） In the formula: , These are the electricity sales revenue and electricity purchase cost for the power distribution network operator. Price in response to the needs of the park; ； （14） In the formula: The price at which distribution network operators purchase electricity from the superior power grid; The optimized system electricity price; This refers to the power consumption of the park after demand response; , These are the compensation coefficients for energy that can be reduced and energy that can be transferred, respectively. , These refer to the power that can be reduced and the power that can be transferred, respectively. The constraints of the upper layer of the two-layer optimization model are: ； （15） In the formula: , , These represent the minimum, maximum, and average electricity sales prices of power distribution network operators, respectively. The lower layer of the two-layer optimization model aims to minimize the overall electricity purchase cost, simultaneously optimizing the revenue of the energy storage operator and the overall energy cost of the park's integrated energy system based on electricity prices and incentive signals. ；（16） ； （17） In the formula, , These are the electricity purchase costs from the power distribution network operator and the integrated energy system of the park, respectively. The constraints of the lower level of the two-level optimization model include: Demand response constraints: ；（18） Power balance constraints: ；（19）。 10. The power system multi-entity collaborative optimization scheduling method according to claim 9, in step 2, the intraday stage collaborative optimization model performs rolling optimization with the objective of minimizing the comprehensive cost of the energy storage operator and the park's integrated energy system within the current rolling time domain, to obtain the intraday optimized scheduling scheme, the objective function of which is: ；（20） ；（21） In the formula: , These are the costs of physical energy storage and virtual energy storage within the day, respectively. , , , These are the gas purchase cost, carbon emission cost, electricity purchase cost, and operation and maintenance cost of the park's integrated energy system within the day; , These are the intraday charging price for electric vehicles and the electricity sales price for energy storage. The electricity sales price for the park's integrated energy system; This refers to the daily gas purchase volume of the system. , These are the charging and discharging power of the intraday thermal storage device, , These are the charging and discharging power of the cold storage device, respectively. Energy conversion equipment Daily output power.

Images