A Virtual Power Plant Optimization Operation Method Based on Master-Slave Game Theory in Rural Scenario

Abstract

A method for optimizing the operation of a virtual power plant in a rural setting based on a master-slave game theory approach includes: acquiring virtual power plant parameter information; establishing a virtual power plant model, including an OLTC model, an air conditioning model, an electric vehicle model, a photovoltaic system model, a static var compensator (SVC) model, and a battery model; establishing a V2G game theory model for electric bicycles, using an evolutionary game with the Logit criterion and maximizing user revenue as the objective function; establishing a two-layer game theory model with economic benefits and curtailment rate as objectives, consisting of a virtual power plant model at the upper layer and an electric bicycle V2G model at the lower layer; solving the two-layer game theory model to obtain the charging and discharging strategies of individual energy storage components, the charging and discharging strategies of electric bicycles, the voltage and power of each node, user economic benefits, and the system curtailment rate, thereby obtaining the optimal operation strategy. This invention improves the environmental and economic benefits of rural power distribution networks by constructing a master-slave game theory model with the objectives of maximizing economic benefits and minimizing curtailment rate.

Description

Technical Field

[0001] This invention relates to the field of power grid operation optimization, and in particular to a virtual power plant operation optimization method based on master-slave game theory in a rural setting. Background Technology

[0002] In response to the high carbon emissions and pollution of traditional power systems, new power systems based on distributed energy sources such as wind power and solar power have received widespread attention. However, the large-scale grid connection of renewable energy sources such as wind power and solar power, along with the decommissioning of numerous thermal power units, will put even greater pressure on the power grid due to its inherent volatility and uncertainty. In severe cases, this could disrupt power system stability and affect users' normal electricity consumption. Faced with the large number of heterogeneous distributed energy sources in new power systems, traditional control methods struggle to meet all the demands, inevitably leading to either overly conservative or overly lenient control, resulting in significant safety hazards and energy waste. Virtual Power Plants (VPPs), as a novel technology, aggregate different types of distributed energy sources in the system to form a stable and controllable energy set, thereby achieving overall dispatch of the power system. Furthermore, due to the diversity of aggregation methods, VPPs can flexibly adjust to different scenarios and standards, providing the power system with a highly flexible and adaptable distributed energy management method.

[0003] Effectively utilizing my country's abundant rooftop resources not only unleashes enormous potential for the development of distributed photovoltaic (PV) power but also significantly improves the economic efficiency of rural power grids. However, due to the low voltage levels and utilization rates of rural distribution networks, and the highly time-dependent nature of rural loads (high load during busy farming seasons and low load during slack seasons), grid connection conditions for rural distribution networks are poor, affecting PV grid integration. Furthermore, the large number of small inverters introduced by the high-penetration intermittent distributed PV grid integration further negatively impacts the operation of rural AC / DC hybrid distribution networks.

[0004] According to statistics from the China Bicycle Association, the number of electric bicycles in China reached nearly 325 million in 2020, and is projected to reach 400 million by 2022. Most of these are located in rural areas and third- and fourth-tier cities, while air conditioning accounts for more than one-third of electricity consumption during peak hours. The disorderly connection of a large number of flexible loads further exacerbates the burden on rural power distribution networks. However, current research on optimizing the operation and improving the stability of rural power distribution networks is limited, and the characteristics of their loads are not adequately explored.

[0005] With the rapid development of the Internet of Things (IoT) and mobile communications, demand-side management of the power system will become the mainstream in the future. By providing economic compensation to users, flexible loads will proactively and orderly participate in grid regulation. Among these technologies, vehicle-to-grid (V2G) technology gives the grid greater flexibility, enabling it to accommodate more electric vehicle loads and renewable energy output in a more economical, safe, and stable manner. In the long term, V2G effectively promotes the transition of my country's energy structure from fossil fuels to renewable energy, contributing to the realization of the carbon peak and carbon neutrality vision. From the perspective of the electric vehicle industry's development, the implementation of V2G brings considerable returns to participating vehicle owners, increasing their willingness to purchase electric vehicles and indirectly promoting the popularization of electric vehicles and industry growth.

[0006] In summary, there is currently limited research on the coordinated operation of multiple loads in rural scenarios. Under the broader context of "county-wide photovoltaic projects," demand-side management strategies for rural power distribution networks have significant research value and application prospects. Summary of the Invention

[0007] The purpose of this invention is to provide a virtual power plant optimization operation method based on master-slave game theory in rural scenarios. Based on the virtual power plant, a master-slave game model is constructed to maximize economic benefits and minimize curtailment rate, thereby improving the environmental and economic benefits of rural power distribution networks.

[0008] To achieve the above objectives, the present invention provides the following solution:

[0009] A method for optimizing the operation of a virtual power plant in a rural setting based on master-slave game theory includes the following steps:

[0010] Step 1: Obtain virtual power plant parameter information, which includes electric bicycle operating parameters, photovoltaic output, inverter parameters, air conditioner operating parameters, battery operating parameters, static var compensator operating parameters, on-load tap changer parameters, and various load information.

[0011] Step 2: Establish a virtual power plant model based on the virtual power plant parameter information. The virtual power plant model includes an OLTC model, a battery model, a photovoltaic system model, an inverter model, an air conditioning model, an electric vehicle model, a static var compensator, and an AC / DC steady-state power flow model containing AC and DC systems.

[0012] Step 3: Establish a V2G game model for electric bicycles. The V2G game model for electric bicycles adopts an evolutionary game based on the Logit criterion, with the objective function being the maximization of user revenue.

[0013] Step 4: Based on the virtual power plant model established in Step 2 and the electric bicycle V2G game model established in Step 3, establish a two-layer master-slave game model. The upper-layer model is the virtual power plant model, which provides the lower-layer model with electricity pricing strategies based on economic benefits and curtailment rate. The lower-layer model is the electric bicycle V2G game model, which generates output strategies and reserve capacity based on the electricity pricing model provided by the upper-layer model, thus affecting the output of the upper-layer model.

[0014] Step 5: Solve the two-layer master-slave game model to obtain the charging and discharging strategies of each energy storage element, the charging and discharging strategy of the electric bicycle, the voltage and power of each node, the economic benefits to users, and the system curtailment rate, thereby obtaining a set of optimal operating strategies.

[0015] Furthermore, the parameter information of the virtual power plant in step one is as follows:

[0016] a. Operating parameters of electric bicycles: charging and discharging power, bicycle battery loss coefficient, and user travel patterns;

[0017] b. Photovoltaic output: Maximum photovoltaic output;

[0018] c. Inverter parameters: inverter power factor angle, control angle, For commutation reactor, This refers to the inverter voltage turns ratio;

[0019] d. Air conditioning operating parameters: indoor temperature, room equivalent heat capacity, room equivalent thermal resistance, load on / off status, air conditioning cooling / heating power, rated power of a single air conditioner, air conditioning temperature setpoint. This refers to the temperature dead zone bandwidth, the lower and upper limits of the indoor temperature when the air conditioner is switched on and off;

[0020] e. Battery operating parameters: charging and discharging power of energy storage unit, minimum and maximum rated capacity limits of energy storage unit, and charging and discharging efficiency of battery;

[0021] f. Static Var Compensator Operating Parameters: Maximum Output of Static Var Compensator;

[0022] e. On-load tap-changing transformer parameters: regulating voltage of the on-load tap-changing transformer, maximum number of operations per day of the on-load tap-changing transformer;

[0023] g. Various load information: residential load, agricultural production load.

[0024] Furthermore, the virtual power plant model in step two specifically includes:

[0025] a. OLTC model

[0026] When the system experiences voltage fluctuations due to disturbances or partial faults, the on-load tap-changing transformer (OLTC) adjusts the root node voltage by changing the tap position, thereby ensuring that the system voltage level remains normal and achieving optimal power dispatching for the entire system. Its mathematical model is as follows:

[0027] (1)

[0028] In the formula: for node Voltage at any given moment; The regulating voltage for OLTC; To adjust the Boolean coefficient of the voltage, when When the value is 1, OLTC takes action; for The Boolean coefficients of OLTC at time 1, when When it is 1, it means that OLTC is in Momentary action; This represents the maximum number of actions an OLTC can perform per day.

[0029] b. Battery model

[0030] The battery model mainly considers the charging and discharging power limit, operating state constraints, capacity update limitations, and reactive power output constraints. Its mathematical model is as follows:

[0031]

[0032] Where, For the Each energy storage unit in The amount of charge at any given moment. For the Each energy storage unit in The amount of discharge at any given moment; In charging working state, It is in discharge working state; Harm The first Minimum and maximum rated capacity limits for energy storage units; For the Each energy storage unit in The reactive power output value at any given moment. For the Apparent power of each energy storage unit; For battery charging efficiency, For discharge efficiency; For the The remaining power of each energy storage unit at the initial moment. For the The remaining power of each energy storage unit at the time of termination;

[0033] c. Photovoltaic system model

[0034] The output model of a photovoltaic generator is as follows:

[0035]

[0036] for Photovoltaic output power at all times; This refers to the rated output power of the photovoltaic system. and They are respectively rated light radiation and Constant light radiation; and Rated temperature and Temperature at any given time; The power temperature coefficient;

[0037] The light intensity follows a Beta distribution, with the following probability distribution:

[0038]

[0039] and These are the shape parameters of the Beta distribution;

[0040] d. Inverter Model

[0041] Inverter DC voltage and inverter AC bus voltage And the following relationship exists between other parameters:

[0042] (5)

[0043] Where, , Let be constants, and their values ​​are respectively , ; For control angle; For commutation reactance; This refers to the inverter voltage turns ratio;

[0044] DC current With secondary fundamental current The relation is:

[0045]

[0046] Where, It is a constant, usually taken as This will satisfy the calculation accuracy requirements;

[0047] Neglecting losses in the transformer, the AC active power of the converter should be equal to the DC power, that is:

[0048]

[0049] Where, The inverter power factor angle; This refers to the primary current of the converter transformer.

[0050] Consider the following five control methods for the inverter:

[0051] Constant current control:

[0052]

[0053] Constant voltage control:

[0054]

[0055] Constant power control:

[0056]

[0057] Constant power angle control:

[0058]

[0059] Transformer turns ratio control:

[0060]

[0061] e. Air Conditioning Model

[0062] The thermal effect of air conditioning is modeled using the internationally accepted thermal equivalence model, as follows:

[0063]

[0064]

[0065]

[0066] In the formula: for The indoor temperature at any given time; Outdoor temperature; The equivalent heat capacity of the room; The equivalent thermal resistance of the room; This indicates the on / off state of the load. This refers to the air conditioner's cooling / heating power, where... Energy efficiency ratio This refers to the rated power of a single air conditioner. Set the air conditioner temperature setting; This refers to the bandwidth of the temperature dead zone. and These are the lower and upper limits of the indoor temperature when the air conditioner is switched on and off, respectively. It is an infinite time delay;

[0067] Discretizing the above differential equation yields the following expression:

[0068]

[0069] Considering the air conditioning load and indoor temperature, set temperature constraints and power constraints:

[0070]

[0071] In the formula express Space No. Temperature at any moment; , These represent the maximum and minimum indoor temperatures, respectively. Indicates the first air conditioner The workload of the moment; , This indicates the maximum and minimum output power of the air conditioner;

[0072] f. Electric bicycle model

[0073] The load model for electric bicycles includes:

[0074]

[0075] Where: These are Boolean variables representing the charging and discharging of an electric bicycle, respectively. When the electric bicycle discharges at time 1, Charge the electric bicycle at 1 hour; and These are the charging power and discharging power of the electric bicycle, respectively. and The maximum power for charging and discharging electric bicycles;

[0076] g. AC / DC steady-state power flow model

[0077] The steady-state power flow model is the DistFlow power flow model, as detailed below:

[0078]

[0079] In the formula: It indicates three phases: day before, intraday, and real-time; A scheduling period is represented as The scheduling cycles for all three stages are the same. Random scenarios that generate power for photovoltaic systems; and Branch roads Active power and reactive power; and Branch roads Active power and reactive power; and Injection nodes Active power and reactive power; and These refer to the active and reactive power outputs of distributed generation sources in the AC power grid. and They are nodes Active and reactive loads; and These are the active power output and reactive power output at the feeder head end, respectively. Reactive power compensation for the converter; Inject equivalent active power into the AC side nodes of the converter; Increase in electricity purchases from the main grid; Reduce the amount of electricity purchased from the main grid; The flexible load response in this paper is a real-time stage load. The flexible load in this paper is a directly interrupted load, which is essentially a power injection into the node. The first node in the AC power grid is The set of end nodes of the branches; For the end node in the AC power grid The set of the first nodes of the branches; A set of communication nodes; and Branch resistance; and For branch circuit reactance;

[0080] Based on the AC power flow model, a DC system power flow model is constructed as follows:

[0081]

[0082] In the formula: The first node in the DC power grid is The set of end nodes of the branches; The last node in the DC power grid is The set of the first nodes of the branches; A set of DC nodes; A collection of DC branches; and These represent the charging and discharging power of the day-ahead energy storage device; It provides active power output for distributed photovoltaic power generation in DC grids.

[0083] Furthermore, step three, establishing the V2G game model for electric bicycles, specifically includes:

[0084] First, establish the strategy set for evolutionary game, as follows:

[0085]

[0086] In the formula: For the A set of charging strategies for bicycles; Represents the set of elements. Individual charging and discharging strategies; For the Each strategy corresponds to The charging and discharging power at any given time, if Then, when charging an electric bicycle, if Then the electric bicycle discharges.

[0087] Correspondingly, a payment function is constructed with the goal of minimizing the payment cost for electric bicycle users:

[0088]

[0089] In the formula: when hour, , It is a constant, the magnitude of which is the system electricity price; when hour, , To compensate for the discharge, the electricity price is lower when the system is at its peak and higher when the system is at its trough, thereby achieving orderly discharge of electric bicycles and achieving the purpose of peak shaving and valley filling.

[0090] The overall energy consumption coefficient for charging and discharging is expressed as follows:

[0091]

[0092] In the formula: For the cost of battery degradation, The overall benefits of contributing power to the system for electric bicycles. The larger the value, the more beneficial the electric bicycle load is to the system; , This is the comprehensive benefit coefficient; Battery rated capacity; Battery replacement costs; The comprehensive energy consumption coefficient; For the vehicle The state of charge at any given moment;

[0093] Furthermore, the strategy set must satisfy the following constraints:

[0094] Maximum charge / discharge constraints:

[0095]

[0096] In the formula: These represent the maximum and minimum values ​​of the charging and discharging power, respectively.

[0097] Battery charge / discharge equation constraints:

[0098]

[0099] Battery state of charge constraints:

[0100]

[0101] In the formula: These represent the minimum and maximum values ​​of the state of charge, respectively;

[0102] The dynamic evolution equations are constructed using the Logit protocol based on the generated policy set, as follows:

[0103] First, construct the fitness function for each strategy, as follows:

[0104] (27)

[0105] In the formula: For the fitness of each strategy; The probability of choosing each strategy in the population;

[0106] Based on the fitness functions of each strategy, the dynamic evolution equation of the electric bicycle is obtained using the Logit protocol:

[0107]

[0108] That is, from the strategy To strategy The conditional transition probability; Let the noise level be denoted as ; after discretizing the equation, the dynamic evolution equation of the electric bicycle is obtained as follows:

[0109] (29)

[0110] Finally, the optimal charging and discharging power of the electric bicycle under each operating condition is obtained, and the reserve power of the electric bicycle is calculated using the following formula:

[0111]

[0112] and The maximum charging power and maximum discharging power of the electric bicycle.

[0113] Furthermore, step four, which establishes a two-layer master-slave game model, specifically includes:

[0114] Using the virtual power plant model as the upper-level model, and taking the minimum photovoltaic light efficiency and the minimum system power dispatch cost as the objective functions, the objective function is as follows:

[0115]

[0116] Where: This represents the total daily dispatch cost of a virtual power plant. and These are the compensation electricity prices for the output and load of electric bicycles, respectively. and They are respectively The load that a virtual power plant sells and buys from the grid at any given moment; and They are The prices at which a virtual power plant sells and buys electricity from the grid at any given moment; and They are respectively Real-time active power output and load values ​​of distributed energy storage; and They are respectively Compensation price and purchase price for distributed energy storage output at all times; yes The cost of curtailing solar power in real time; They are respectively The maximum and actual values ​​of photovoltaic output at any given time;

[0117] Furthermore, the objective function satisfies the electricity purchase and sale constraints and the electricity price constraints, as follows:

[0118]

[0119] Where: The first equation represents the non-transferable load of the system; the second equation ensures that both the user and the power grid will benefit.

[0120] Using the V2G game model of electric bicycles as the lower-level model, and taking the maximization of user economic benefits as the objective function, the objective function is obtained as follows:

[0121] (33).

[0122] Furthermore, solving the two-layer master-slave game model in step five specifically includes:

[0123] 1) Input virtual power plant parameter information, including electric vehicles, photovoltaic systems, air conditioning systems, batteries, static var compensators (SVCs), OLTC parameter economic costs, carbon emissions, safe operation constraints, and various load information;

[0124] 2) Initialize the population and generate initial electricity price information for electric bicycle users and initial exchange power information for the power grid sent by the virtual power plant;

[0125] 3) The lower-level optimization model is invoked to calculate the charging and discharging reserve capacity of the bicycle battery using the electricity price information of the virtual power plant;

[0126] 4) The upper-level optimization model is invoked to substitute the charging and discharging reserve capacity of electric bicycles into the virtual power plant model;

[0127] 5) Call the subroutine to calculate the individual fitness value and find the current central solution and population similarity. Then, adopt the double mutation strategy to iterate until the iteration limit is reached to obtain the optimal operation of the virtual power plant under the current state and the power purchase and sale scheme, and send it into the lower layer model.

[0128] 6) Check if the number of iterations has reached the upper limit. If so, output the optimal scheduling strategy; otherwise, update the particles and return to step 3.

[0129] In summary, the above-described technical solutions conceived in this invention can achieve the following beneficial effects:

[0130] 1. A virtual power plant model was constructed, which comprehensively considered various controllable loads within the power plant, fully explored their coordination and complementarity capabilities, improved the system's economic efficiency, mitigated curtailment of solar power, and optimized system operation configuration.

[0131] 2. The evolutionary game model fully considers electric bicycles and the high flexibility and bounded rationality of bicycle users, thus highly replicating the real situation.

[0132] 3. By adopting a master-slave game model, the interaction between electric bicycles and virtual power plants is considered, thus balancing the interests of all parties and improving the economic efficiency and environmental friendliness of the system operation. Attached Figure Description

[0133] Figure 1 This invention provides an embodiment of a virtual power plant optimization method based on master-slave game theory in a rural setting.

[0134] Flowchart of the operation method;

[0135] Figure 2 This is a schematic diagram of the integrated virtual power plant structure of the present invention;

[0136] Figure 3 This is a flowchart of the comprehensive evolutionary game theory process of this invention;

[0137] Figure 4 This is a flowchart illustrating the master-slave game theory process of this invention.

[0138] Figure 5 To improve the IEEE 33-node topology;

[0139] Figure 6 A trend chart showing the proportion of each strategy in an evolutionary game;

[0140] Figure 7 Contributing to EBA;

[0141] Figure 8 This is a diagram showing the power output of each load in the power system.

[0142] Figure 9 This refers to the electricity price charged by the VPP to the grid and the electricity price charged by the VPP for exchanging electricity with the EBA.

[0143] intention. Detailed Implementation

[0144] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. However, the scope of protection of the present invention is not limited to the embodiments described.

[0145] like Figure 1 As shown, this embodiment of the invention provides a method for optimizing the operation of a virtual power plant based on master-slave game theory in a rural setting, including the following steps:

[0146] 1) Obtain virtual power plant parameter information, including electric vehicles, photovoltaic systems, air conditioning systems, batteries, static var compensators (SVCs), OLTC parameter economic costs, carbon emissions, safe operation constraints, and various load information;

[0147] 2) Based on the virtual power plant parameter information, establish a virtual power plant model, including an OLTC model, an air conditioning model, an electric vehicle model, a photovoltaic system model, a static var compensator (SVC), a battery model, and an AC / DC steady-state power flow model. The photovoltaic system model includes solar panels and inverter devices, such as... Figure 2 As shown;

[0148] A. OLTC Model

[0149] An on-load tap-changing transformer (OLTC) can regulate the root node voltage by adjusting the tap position when the system experiences voltage fluctuations due to disturbances or partial faults, thereby ensuring that the system voltage level remains normal and achieving optimal power dispatching for the entire system. Its mathematical model is as follows:

[0150] (1)

[0151] Where: for node Voltage at any given moment; The regulating voltage for the OLTC; To adjust the Boolean coefficient of the voltage, when When the value is 1, OLTC takes action; for The Boolean coefficients of OLTC at time 1, when When it is 1, it means that OLTC is in Momentary action; This represents the maximum number of actions OLTC can perform per day.

[0152] B. Battery Model

[0153] The model of an electric energy storage battery system mainly considers several aspects, including charge and discharge power limits, operating state constraints, capacity update limitations, and reactive power output constraints.

[0154] (2)

[0155] In the formula, For the first Each energy storage unit in The amount of charge at any given moment. For the first Each energy storage unit in The amount of discharge at any given moment; In charging working state, It is in discharge working state; Harm The first Minimum and maximum rated capacity limits for energy storage units; For the Each energy storage unit in The reactive power output value at any given moment. For the Apparent power of each energy storage unit; For battery charging efficiency, For discharge efficiency; For the The remaining power of each energy storage unit at the initial moment. For the The remaining power of each energy storage unit at the time of termination.

[0156] C. Solar cell model

[0157] The output model of a photovoltaic generator is as follows:

[0158] (3)

[0159] for Photovoltaic output power at all times; This refers to the rated output power of the photovoltaic system. and They are respectively rated light radiation and Constant light radiation; and Rated temperature and Temperature at any given time; This is the power temperature coefficient.

[0160] Regarding the stochastic nature of photovoltaics, existing research has shown that, under normal circumstances, light intensity follows a Beta distribution, with the following probability distribution:

[0161] (4)

[0162] and These are the shape parameters of the Beta distribution.

[0163] D. Inverter Model

[0164] Inverter DC voltage and inverter AC bus voltage And the following relationship exists between other parameters:

[0165] (5)

[0166] In the formula, , Let be constants, and their values ​​are respectively , ; For control angle; For commutation reactance; This refers to the inverter voltage transformation ratio.

[0167] DC current With secondary fundamental current The relation is:

[0168] (6)

[0169] In the formula, It is a constant, usually taken as This will satisfy the required calculation accuracy.

[0170] Neglecting losses in the transformer, the AC active power of the converter should be equal to the DC power, that is:

[0171] (7)

[0172] In the formula, The inverter power factor angle; This refers to the primary current of the converter transformer.

[0173] Furthermore, to adapt to different operational requirements, the following five control methods are typically considered for inverters:

[0174] Constant current control:

[0175] (8)

[0176] Constant voltage control:

[0177] (9)

[0178] Constant power control:

[0179] (10)

[0180] Constant power angle control:

[0181] (11)

[0182] Transformer turns ratio control:

[0183] (12)

[0184] E. Air Conditioner Model

[0185] The thermal effect of air conditioning is modeled using the internationally accepted equivalent thermal parameter (ETP) model, as follows:

[0186] (13)

[0187] (14)

[0188] (15)

[0189] Where: for The indoor temperature at any given time; Outdoor temperature; The equivalent heat capacity of the room; The equivalent thermal resistance of the room; This indicates the on / off state of the load. This refers to the air conditioner's cooling / heating power, where... Energy efficiency ratio This refers to the rated power of a single air conditioner. Set the air conditioner temperature setting; This refers to the bandwidth of the temperature dead zone. and These are the lower and upper limits of the indoor temperature when the air conditioner is switched on and off, respectively. It is an infinite time delay (set as the simulation time step in the case of discrete-time simulation).

[0190] Discretizing the above differential equation yields the following expression:

[0191] (16)

[0192] Furthermore, considering the air conditioning load and indoor temperature, temperature constraints and power constraints also need to be set:

[0193] (17)

[0194] In the formula express Space No. Temperature at any moment; , These represent the maximum and minimum indoor temperatures, respectively. Indicates the first air conditioner The workload of the moment; , This indicates the maximum and minimum output power of the air conditioner.

[0195] F. Electric bicycle model

[0196] The load model for electric bicycles includes the following aspects:

[0197] (18)

[0198] Where: These are Boolean variables representing the charging and discharging of an electric bicycle, respectively. When the electric bicycle discharges at time 1, Charge the electric bicycle at 1 hour; and These are the charging power and discharging power of the electric bicycle, respectively. and The maximum power for charging and discharging an electric bicycle.

[0199] G. AC / DC steady-state power flow model

[0200] The DistFlow power flow model for AC / DC distribution networks is as follows:

[0201] (19)

[0202] In the formula: It indicates three phases: day before, intraday, and real-time; A scheduling period is represented as The scheduling cycles for all three stages are the same. Random scenarios that generate power for photovoltaic systems; and Branch roads Active power and reactive power; and Branch roads Active power and reactive power; and Injection nodes Active power and reactive power; and These refer to the active and reactive power outputs of distributed generation sources in the AC power grid. and They are nodes Active and reactive loads; and These are the active power output and reactive power output at the feeder head end, respectively. Reactive power compensation for the converter; Inject equivalent active power into the AC side nodes of the converter; Increase in electricity purchases from the main grid; Reduce the amount of electricity purchased from the main grid; The flexible load response in this paper is a real-time load, which is a direct interruption load, essentially a power injection into the node. The first node in the AC power grid is The set of end nodes of the branches; For the end node in the AC power grid The set of the first nodes of the branches; A set of communication nodes; and Branch resistance; and For branch circuit reactance.

[0203] Next, based on the AC power flow model, we construct the DC system power flow model, as follows:

[0204] (20)

[0205] In the formula: The first node in the DC power grid is The set of end nodes of the branches; The last node in the DC power grid is The set of the first nodes of the branches; A set of DC nodes; A collection of DC branches; and These represent the charging and discharging power of the day-ahead energy storage device; It provides active power output for distributed photovoltaic power generation in DC grids.

[0206] 3) Establish a V2G game model for electric bicycles, using evolutionary game theory based on the Logit criterion, with the objective function being maximizing user revenue, such as... Figure 3 As shown.

[0207] Because the charging and discharging behavior of electric bicycles is highly random and exhibits bounded rationality, and evolutionary game theory, which is based on the assumption of bounded rationality and uses dynamic processes to study how participants adjust their behavior to adapt to the environment or opponents during game evolution, thereby generating trends in group behavior evolution, using evolutionary game theory to describe the charging and discharging behavior of bicycles is closer to reality.

[0208] First, establish the strategy set for evolutionary game, as follows:

[0209] (twenty one)

[0210] Where: For the first A set of charging strategies for bicycles; Represents the set of elements. Individual charging and discharging strategies; For the Each strategy corresponds to The charging and discharging power at any given time, if Then, when charging an electric bicycle, if Then the electric bicycle discharges.

[0211] Correspondingly, a payment function is constructed with the goal of minimizing the payment cost for electric bicycle users:

[0212] (twenty two)

[0213] In the formula: when hour, , It is a constant, the magnitude of which is the system electricity price; when hour, , To compensate for the discharge, the electricity price is lower when the system is at its peak and higher when the system is at its trough. This allows for the orderly discharge of electric bicycles, thus achieving the purpose of peak shaving and valley filling.

[0214] and The overall energy consumption coefficient for charging and discharging is expressed as follows:

[0215] (twenty three)

[0216] In the formula: For the cost of battery degradation, The overall benefits of contributing power to the system for electric bicycles. The larger the value, the more beneficial the electric bicycle load is to the system; , This is the comprehensive benefit coefficient; Battery rated capacity; Battery replacement costs; The comprehensive energy consumption coefficient; For the vehicle The state of charge at any given moment.

[0217] Furthermore, the strategy set must satisfy the following constraints:

[0218] Maximum charge and discharge constraints

[0219] (twenty four)

[0220] In the formula: These represent the maximum and minimum values ​​of the charging and discharging power, respectively.

[0221] Battery charge / discharge equation constraints:

[0222] (25)

[0223] Battery state of charge constraints:

[0224] (26)

[0225] In the formula: These represent the minimum and maximum values ​​of the state of charge, respectively.

[0226] Next, based on the generated policy set, a dynamic evolution equation is constructed using the Logit protocol, as follows:

[0227] First, construct the fitness function for each strategy, as follows:

[0228] (27)

[0229] In the formula: For the fitness of each strategy; The probability of choosing each strategy in the population.

[0230] Based on the fitness of each strategy, the dynamic evolution equation of the electric bicycle can be obtained using the Logit protocol as follows:

[0231] (28)

[0232] That is, from the strategy To strategy The conditional transition probability; Let be the noise level. After discretizing the equation, the dynamic evolution equation of the electric bicycle is obtained as follows:

[0233] (29)

[0234] Finally, the optimal charging and discharging power of the electric bicycle under each operating condition can be obtained, and the reserve power of the electric bicycle can be calculated using the following formula:

[0235] (30)

[0236] and The maximum charging power and maximum discharging power of the electric bicycle.

[0237] 4) With economic benefits and curtailment rate as objectives, a two-layer master-slave game model is established, with the upper layer being a virtual power plant model and the lower layer being an electric bicycle V2G model.

[0238] Using the virtual power plant model as the upper-level optimization model, and taking the minimization of photovoltaic power efficiency and the minimization of system power dispatch cost as the objective functions, the objective function is as follows:

[0239] (31)

[0240] In the formula: This represents the total daily dispatch cost of a virtual power plant. and These are the compensation electricity prices for the output and load of electric bicycles, respectively. and They are respectively The load that a virtual power plant sells and buys from the grid at any given moment; and They are The prices at which a virtual power plant sells and buys electricity from the grid at any given moment; and They are respectively Real-time active power output and load values ​​of distributed energy storage; and They are respectively Compensation price and purchase price for distributed energy storage output at all times; yes The cost of curtailing solar power in real time; They are respectively The maximum and actual values ​​of photovoltaic power output at any given time.

[0241] Furthermore, the objective function must also satisfy the electricity purchase and sale constraints and the electricity price constraints, as follows:

[0242] (32)

[0243] In the formula: The first equation represents the non-transferable load of the system; the second equation ensures that both the user and the power grid will benefit.

[0244] Using the evolutionary game model of electric bicycles as the lower-level optimization model, and taking the maximization of user economic benefits as the objective function, the objective function is obtained as follows:

[0245] (33)

[0246] Based on the upper and lower layer optimization model, the master-slave game relationship can be obtained as follows:

[0247] 1. The virtual power plant is the leader. Based on the market clearing results and information such as energy storage, load and distributed renewable energy output collected by the Internet of Things, it selects a pricing strategy from the upper-level strategy space (VPP) and releases it to users.

[0248] 2. Electric bicycle users are followers, adjusting their energy consumption within a certain range and selecting the optimal response strategy from the lower-level strategy space based on the published price.

[0249] 3. The virtual power plant obtains the optimal pricing strategy based on the response strategy of electric bicycle users, and repeats the above game process until equilibrium is reached.

[0250] 5) An improved double-mutation differential evolution algorithm is proposed to solve the two-layer master-slave game model, obtaining the charging and discharging strategies of individual energy storage elements, the charging and discharging strategies of electric bicycles, the voltage and power of each node, the economic benefits to users, and the system curtailment rate, thereby obtaining a set of optimal operating strategies, such as... Figure 5 As shown.

[0251] The algorithm improvements are as follows:

[0252] 1) Population similarity index and central solution

[0253] The definitions of population similarity index and central solution in the improved double-mutation differential evolution algorithm are as follows:

[0254] (34)

[0255] Where: The central solution for generation g; Population size; Individuals of generation g; The similarity of the g-generation population; , and These represent the average fitness value of the g-generation population, the fitness value of the optimal solution, and the fitness value of the worst solution, respectively.

[0256] 2) The double mutation strategy is:

[0257] (35)

[0258] Where: Individuals that are variants in the g generation population; , and for Not equal to Two pairwise distinct integers; This is a local parameter, set to 0.1; rand indicates that in Random numbers within an interval; The scaling factor is typically a fixed value in traditional differential evolution algorithms, which limits the algorithm's search capability to some extent. Therefore, this paper... Pick Random numbers within an interval.

[0259] The algorithm's solution steps are as follows:

[0260] 1) Input virtual power plant parameter information, including electric vehicles, photovoltaic systems, air conditioning systems, batteries, static var compensators (SVCs), OLTC parameter economic costs, carbon emissions, safe operation constraints, and various load information;

[0261] 2) Initialize the population and generate the initial electricity price information sent by the virtual power plant to electric bicycle users and the initial exchange power information sent to the power grid.

[0262] 3) Call the lower-level optimization model to calculate the charging and discharging reserve capacity of the bicycle battery using the electricity price information of the virtual power plant.

[0263] 4) Call the upper-level optimization model to substitute the charging and discharging reserve capacity of electric bicycles into the virtual power plant model.

[0264] 5) Call the subroutine to calculate the individual fitness value and find the current central solution and population similarity. Then, adopt the double mutation strategy to iterate until the iteration limit is reached to obtain the optimal operation of the virtual power plant under the current state and the power purchase and sale scheme, and send it into the lower layer model.

[0265] 6) Check if the number of iterations has reached the upper limit. If so, output the optimal scheduling strategy; otherwise, update the particles and return to step 3.

[0266] Based on the above method, the following example is constructed for verification. The example parameters are shown in Table 1:

[0267] Table 1 System Parameters

[0268]

[0269] Rural power grids are characterized by low grid levels and long power supply radii. To address this, this invention uses an improved IEEE-33 node system for analysis. Nodes 16, 17, 21, 23, 24, 31, and 32 are DC nodes. Rooftop photovoltaic systems are connected to nodes 17, 21, 24, and 32, distributed energy storage is connected to nodes 12, 17, 21, 24, and 32, and SVC systems are connected to nodes 5, 11, and 29. The system voltage reference value is set to 10kV, and the power reference value is 3MW. The distribution network topology is shown below. Figure 4 As shown.

[0270] Based on the above parameters, and considering the characteristics of rural scenarios, two scenarios were constructed for verification, taking into account the different load distributions during the off-season and busy season. It is assumed that the system contains 20 EBAs, and each EBA contains 100 EB users. The final evolution trend is as follows: Figure 6 As shown:

[0271] At the same time, the output of EBA in different scenarios can be obtained, such as Figure 7As shown in Table 2, the final energy performance indicators for each scenario, considering and not considering EBA, are as follows:

[0272] Table 2 Electricity Indicators

[0273]

[0274] Based on this, the output values ​​of each load in the power system during the off-season and busy seasons are obtained as follows: Figure 8 As shown, the electricity price charged by the VPP to the grid and the electricity price charged by the VPP for exchanging electricity with the EBA are as follows: Figure 9 As shown.

[0275] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for optimizing the operation of a virtual power plant in a rural setting based on master-slave game theory, characterized in that, Includes the following steps: Step 1: Obtain virtual power plant parameter information, which includes electric bicycle operating parameters, photovoltaic output, inverter parameters, air conditioner operating parameters, battery operating parameters, static var compensator operating parameters, on-load tap changer parameters, and various load information. Step 2: Establish a virtual power plant model based on the virtual power plant parameter information. The virtual power plant model includes an OLTC model, a battery model, a photovoltaic system model, an inverter model, an air conditioning model, an electric vehicle model, a static var compensator, and an AC / DC steady-state power flow model containing AC and DC systems. Step 3: Establish a V2G game model for electric bicycles. The V2G game model for electric bicycles adopts an evolutionary game based on the Logit criterion, with the objective function being the maximization of user revenue. Step 4: Based on the virtual power plant model established in Step 2 and the electric bicycle V2G game model established in Step 3, establish a two-layer master-slave game model. The upper-layer model is the virtual power plant model, which provides the lower-layer model with electricity pricing strategies based on economic benefits and curtailment rate. The lower-layer model is the electric bicycle V2G game model, which generates output strategies and reserve capacity based on the electricity pricing strategies provided by the upper-layer model, thus affecting the output of the upper-layer model. Step 5: Solve the two-layer master-slave game model to obtain the charging and discharging strategies of each energy storage element, the charging and discharging strategy of the electric bicycle, the voltage and power of each node, the economic benefits to users, and the system curtailment rate, thereby obtaining a set of optimal operating strategies; Step four, establishing a two-layer master-slave game model, specifically includes: Using the virtual power plant model as the upper-level model, and taking the minimization of photovoltaic curtailment rate and system power dispatch cost as the objective functions, the objective function is as follows: （31）； In the formula: This represents the total daily dispatch cost of a virtual power plant. and These are the compensation electricity prices for the output and load of electric bicycles, respectively. and They are respectively The load that a virtual power plant sells and buys from the grid at any given moment; and They are The prices at which a virtual power plant sells and buys electricity from the grid at any given moment; and They are respectively Real-time active power output and load values ​​of distributed energy storage; and They are respectively Compensation price and purchase price for distributed energy storage output at all times; yes The cost of curtailing solar power in real time; They are respectively The maximum and actual values ​​of photovoltaic power output at any given time; and These are the charging power and discharging power of the electric bicycle, respectively. Furthermore, the objective function satisfies the electricity purchase and sale constraints and the electricity price constraints, as follows: ； In the formula: The first equation represents the non-transferable load of the system; the second equation ensures that both the user and the power grid will benefit. Using the V2G game model of electric bicycles as the lower-level model, and taking the maximization of user economic benefits as the objective function, the objective function is obtained as follows: ； In the formula: Represents the set of elements. Individual charging and discharging strategies; For the first Each strategy corresponds to The charging and discharging power at any given moment; when hour, , To compensate for the discharge, the electricity price is lower during system peaks and higher during system troughs, thereby achieving orderly discharge of electric bicycles and realizing peak shaving and valley filling; if Then, when charging an electric bicycle, if Then the electric bicycle discharges. This is the overall energy consumption coefficient for charging and discharging.

2. The method for optimizing the operation of a virtual power plant in a rural setting based on master-slave game theory as described in claim 1, characterized in that, The parameter information of the virtual power plant in step one is as follows: a. Operating parameters of electric bicycles: charging and discharging power, bicycle battery loss coefficient, and user travel patterns; b. Photovoltaic output: Maximum photovoltaic output; c. Inverter parameters: inverter power factor angle, control angle, For commutation reactor, This refers to the inverter voltage ratio; d. Air conditioning operating parameters: indoor temperature, room equivalent heat capacity, room equivalent thermal resistance, load on / off status, air conditioning cooling / heating power, rated power of a single air conditioner, air conditioning temperature setpoint. This refers to the temperature dead zone bandwidth, the lower and upper limits of the indoor temperature when the air conditioner is switched on and off; e. Battery operating parameters: charging and discharging power of energy storage unit, minimum and maximum rated capacity limits of energy storage unit, and charging and discharging efficiency of battery; f. Static Var Compensator Operating Parameters: Maximum Output of Static Var Compensator; e. On-load tap-changing transformer parameters: regulating voltage of the on-load tap-changing transformer, maximum number of operations per day of the on-load tap-changing transformer; g. Various load information: residential load, agricultural production load.

3. The method for optimizing the operation of a virtual power plant in a rural setting based on master-slave game theory, as described in claim 1, is characterized in that... The virtual power plant model in step two specifically includes: a. OLTC model When the system experiences voltage fluctuations due to disturbances or partial faults, the on-load tap-changing transformer (OLTC) adjusts the root node voltage by changing the tap position, thereby ensuring that the system voltage level remains normal and achieving optimal power dispatching for the entire system. Its mathematical model is as follows: ； In the formula: for node Voltage at any given moment; The regulating voltage for the OLTC; To adjust the Boolean coefficient of the voltage, when When the value is 1, OLTC takes action; for The Boolean coefficients of OLTC at time 1, when When it is 1, it means that OLTC is in Momentary action; This represents the maximum number of actions an OLTC can perform per day. b. Battery model The battery model mainly considers the charging and discharging power limit, operating state constraints, capacity update limitations, and reactive power output constraints. Its mathematical model is as follows: ； In the formula, For the first Each energy storage unit in The amount of charge at any given moment. For the first Each energy storage unit in The amount of discharge at any given moment; In charging working state, It is in discharge working state; Harm Respectively Minimum and maximum rated capacity limits for energy storage units; For the first Each energy storage unit in The reactive power output value at any given moment. For the first Apparent power of each energy storage unit; For battery charging efficiency, For discharge efficiency; For the first The remaining power of each energy storage unit at the initial moment. For the first The remaining power of each energy storage unit at the time of termination; c. Photovoltaic system model The output model of a photovoltaic generator is as follows: ； for Photovoltaic output power at all times; This refers to the rated output power of the photovoltaic system. and They are respectively rated light radiation and Constant light radiation; and Rated temperature and Temperature at any given time; The power temperature coefficient; The light intensity follows a Beta distribution, with the following probability distribution: ； and These are the shape parameters of the Beta distribution; d. Inverter Model Inverter DC voltage and inverter AC bus voltage And the following relationship exists between other parameters: (5) ； In the formula, , Let be constants, and their values ​​are respectively , ; For control angle; For commutation reactance; This refers to the inverter voltage ratio; DC current With secondary fundamental current The relation is: ； In the formula, It is a constant, usually taken as This will satisfy the calculation accuracy requirements; Neglecting losses in the transformer, the AC active power of the converter should be equal to the DC power, that is: ； In the formula, The inverter power factor angle; This refers to the primary current of the converter transformer. Consider the following five control methods for the inverter: Constant current control: (8) ； Constant voltage control: ； Constant power control: ； Constant power angle control: ； Transformer turns ratio control: ； e. Air Conditioning Model The thermal effect of air conditioning is modeled using the internationally accepted thermal equivalence model, as follows: ； ； (15)； In the formula: for The indoor temperature at any given time; Outdoor temperature; The equivalent heat capacity of the room; The equivalent thermal resistance of the room; This indicates the on / off state of the load. This refers to the air conditioner's cooling / heating power, where... Energy efficiency ratio This refers to the rated power of a single air conditioner. Set the air conditioner temperature setting; This refers to the bandwidth of the temperature dead zone. and These are the lower and upper limits of the indoor temperature when the air conditioner is switched on and off, respectively. It is an infinite time delay; Discretizing the differential equation yields the following expression: ； Considering the air conditioning load and indoor temperature, set temperature constraints and power constraints: ； In the formula express Space No. Temperature at any moment; , These represent the maximum and minimum indoor temperatures, respectively. Indicates the first air conditioner The workload of the moment; , This indicates the maximum and minimum output power of the air conditioner; f. Electric bicycle model The load model for electric bicycles includes: (18) ； In the formula: These are Boolean variables representing the charging and discharging of an electric bicycle, respectively. When the electric bicycle discharges at time 1, Charge the electric bicycle at 1 hour; and These are the charging power and discharging power of the electric bicycle, respectively. and The maximum power for charging and discharging electric bicycles; g. AC / DC steady-state power flow model The steady-state power flow model is the DistFlow power flow model, as detailed below: ； In the formula: It indicates three phases: previous day, intraday, and real-time; A scheduling period is represented as The scheduling cycles for all three stages are the same. Random scenarios that generate power for photovoltaic systems; and Branch roads Active power and reactive power; and Branch roads Active power and reactive power; and Injection nodes Active power and reactive power; and These refer to the active and reactive power outputs of distributed generation sources in the AC power grid. and They are nodes Active and reactive loads; and These are the active power output and reactive power output at the feeder head end, respectively. Reactive power compensation for the converter; Inject equivalent active power into the AC side nodes of the converter; Increase in electricity purchases from the main grid; Reduce the amount of electricity purchased from the main grid; The flexible load response in this paper is a real-time load, which is a direct interruption load, essentially a power injection into the node. The first node in the AC power grid is The set of end nodes of the branches; For the end node in the AC power grid The set of the first nodes of the branches; A set of communication nodes; and Branch resistance; and For branch circuit reactance; Based on the AC power flow model, a DC system power flow model is constructed as follows: ； In the formula: The first node in the DC power grid is The set of end nodes of the branches; The last node in the DC power grid is The set of the first nodes of the branches; A set of DC nodes; A collection of DC branches; and These represent the charging and discharging power of the day-ahead energy storage device; It provides active power output for distributed photovoltaic power generation in DC grids.

4. The virtual power plant optimization operation method based on master-slave game theory in a rural scenario according to claim 1, characterized in that, Step three, establishing the V2G game model for electric bicycles, specifically includes: First, establish the strategy set for evolutionary game, as follows: (21) ； In the formula: For the first A set of charging strategies for bicycles; Represents the set of elements. Individual charging and discharging strategies; For the first Each strategy corresponds to The charging and discharging power at any given time, if Then, when charging an electric bicycle, if Then the electric bicycle discharges. Correspondingly, a payment function is constructed with the goal of minimizing the payment cost for electric bicycle users: ； In the formula: when hour, , It is a constant, the magnitude of which is the system electricity price; when hour, , To compensate for the discharge, the electricity price is lower when the system is at its peak and higher when the system is at its trough, thereby achieving orderly discharge of electric bicycles and achieving the purpose of peak shaving and valley filling. The overall energy consumption coefficient for charging and discharging is expressed as follows: ； In the formula: For the cost of battery degradation, The overall benefits of contributing power to the system for electric bicycles. The larger the value, the more beneficial the electric bicycle load is to the system; , This is the comprehensive benefit coefficient; Battery rated capacity; Battery replacement costs; The comprehensive energy consumption coefficient; For the first vehicle The state of charge at any given moment; Furthermore, the strategy set must satisfy the following constraints: Maximum charge / discharge constraints: ； In the formula: These represent the maximum and minimum values ​​of the charging and discharging power, respectively. Battery charge / discharge equation constraints: ； Battery state of charge constraints: ； In the formula: These represent the minimum and maximum values ​​of the state of charge, respectively; The dynamic evolution equations are constructed using the Logit protocol based on the generated policy set, as follows: First, construct the fitness function for each strategy, as follows: ； In the formula: For the fitness of each strategy; The probability of choosing each strategy in the population; Based on the fitness functions of each strategy, the dynamic evolution equation of the electric bicycle is obtained using the Logit protocol: ； That is, from the strategy To strategy The conditional transition probability; Let the noise level be denoted as ; after discretizing the equation, the dynamic evolution equation of the electric bicycle is obtained as follows: ； Finally, the optimal charging and discharging power of the electric bicycle under each operating condition is obtained, and the reserve power of the electric bicycle is calculated using the following formula: ； and The maximum charging power and maximum discharging power of the electric bicycle.

5. The virtual power plant optimization operation method based on master-slave game theory in a rural scenario according to claim 1, characterized in that, Step five, solving the two-layer master-slave game model, specifically includes: 1) Input virtual power plant parameter information, including electric vehicles, photovoltaic systems, air conditioning systems, batteries, static var compensators (SVCs), OLTC parameter economic costs, carbon emissions, safe operation constraints, and various load information; 2) Initialize the population and generate initial electricity price information for electric bicycle users and initial exchange power information for the power grid sent by the virtual power plant; 3) The lower-level optimization model is invoked to calculate the charging and discharging reserve capacity of the bicycle battery using the electricity price information of the virtual power plant; 4) The upper-level optimization model is invoked to substitute the charging and discharging reserve capacity of electric bicycles into the virtual power plant model; 5) Call the subroutine to calculate the individual fitness value and find the current central solution and population similarity. Then, adopt the double mutation strategy to iterate until the iteration limit is reached to obtain the optimal operation of the virtual power plant under the current state and the power purchase and sale scheme, and send it into the lower layer model. 6) Check if the number of iterations has reached the upper limit. If so, output the optimal scheduling strategy; otherwise, update the particles and return to step 3.

Images