Firewood-generator-containing virtual power plant control method based on green certificate-stepped carbon bidirectional interaction

Abstract

The invention discloses a firewood-generator-containing virtual power plant control method based on green certificate-stepped carbon bidirectional interaction, and the method comprises the following steps: building a virtual power plant model through the comprehensive consideration of the physical characteristics of all devices in a virtual power plant; a bidirectional interaction mechanism is introduced, and a novel bidirectional interaction model is constructed; based on the virtual power plant model and the bidirectional interaction model, establishing an optimization model with the goal of minimizing the overall operation cost and transaction cost of the virtual power plant, and rewriting the optimization model into a robust optimization model of a min-max-min structure; splitting the robust optimization model into a main problem and a sub-problem by adopting a column constraint generation algorithm and a strong duality theory; camp is adopted; and solving the main problem and the sub-problems by a CG algorithm to obtain a virtual power plant control scheme. According to the method, the renewable energy consumption capability can be effectively improved, and the economy and robustness of the virtual power plant are guaranteed.

Description

Technical Field

[0001] This invention belongs to the field of power systems and relates to virtual power plant optimization technology, specifically to a control method for a diesel-powered virtual power plant based on bidirectional interaction of green certificates and tiered carbon. Background Technology

[0002] With the development of cloud computing, big data analytics, and artificial intelligence, the energy consumption of global data centers has increased significantly, placing enormous pressure on the power grid, especially during peak electricity demand periods. To address this challenge, demand response mechanisms have become a key means of optimizing data center power system access. Simultaneously, to meet the stringent power reliability requirements of data centers, operators equip them with various backup power sources such as energy storage, gas turbines, and diesel generators. Researching unified scheduling of these backup power sources through virtual power plants can effectively improve overall system efficiency. Furthermore, the enormous energy consumption of data centers also leads to carbon emissions. Therefore, to achieve green and sustainable development of data centers and effectively promote the industry's "carbon peaking" and "carbon neutrality" goals, it is urgent to research energy-saving and carbon-reduction optimization methods for data centers, addressing the realities of high energy consumption and high carbon emissions.

[0003] Currently, research on data centers' participation in demand response and grid optimization scheduling has yielded some results: Starting from the load characteristics of data centers, studies have been conducted on load allocation strategies within data centers to adjust load distribution and reduce power costs; an optimized scheduling model for data center power supply systems oriented towards demand response has been established under time-of-use pricing; and an optimization model considering the coordinated operation of energy storage and data centers in microgrids has been established with the goal of minimizing data center operating costs. These studies primarily focus on the direct scheduling of servers or data centers and how to optimize the utilization of the load's inherent portability, but few studies have explored guiding data centers to participate in demand response and virtual power plant optimization scheduling through price compensation mechanisms. Regarding energy conservation and carbon reduction in data centers, existing research includes: promoting renewable energy consumption through renewable energy participation in green certificate trading; constructing a collaborative optimization strategy for the electricity market and green certificate market; and establishing a comprehensive low-carbon economic scheduling model for energy systems that simultaneously considers tiered carbon trading and green certificate trading. However, since green certificates represent the environmental value of green electricity, while carbon emission allowances represent the environmental costs of fossil energy consumption, the interaction between the two inevitably leads to double counting of carbon footprints. Therefore, establishing a carbon-green certificate collaborative trading model based on the carbon emission reduction effect of green certificates remains a key research focus. Summary of the Invention

[0004] Purpose of the invention: In order to overcome the shortcomings of the existing technology, this invention provides a control method for a virtual power plant with diesel generation based on the bidirectional interaction of green certificates and tiered carbon, which can effectively improve the renewable energy consumption capacity and ensure the economic efficiency and robustness of the virtual power plant.

[0005] Technical Solution: To achieve the above objectives, this invention provides a control method for a virtual power plant with diesel generation based on a two-way interaction of green certificates and tiered carbon emissions, comprising the following steps:

[0006] S1: Construct a virtual power plant model by comprehensively considering the physical characteristics of each device in the virtual power plant;

[0007] S2: Introduce a two-way interaction mechanism to construct a new green certificate-step carbon two-way interaction model;

[0008] S3: Based on the virtual power plant model and the two-way interaction model, establish an optimization model with the goal of minimizing the overall operating cost and transaction cost of the virtual power plant, and rewrite it into a robust optimization model with a min-max-min structure.

[0009] S4: Using the column constraint generation algorithm and strong duality theory, the robust optimization model is decomposed into a main problem and sub-problems; the C&CG algorithm is used to solve the main problem and sub-problems to obtain the virtual power plant optimization scheduling scheme.

[0010] Furthermore, the virtual power plant model in step S1 is specifically expressed as follows:

[0011] Data center load model:

[0012] The power model of a data center is expressed as follows:

[0013] (1)

[0014] (2)

[0015] In the formula: This is a variable representing the proportion of the total task load where users convert real-time tasks to batch processing tasks due to price compensation. The number of batch processing tasks. The first report submitted by the user The maximum delay time for each task for The batch processing task load received by the data center at any given moment; for Real-time task load received by the data center at any given moment. and Assigned to The sum of the task load at each moment is the current task load. Total data center workload at any given time ;

[0016] The workload in a data center should be less than the data center's maximum processing capacity.

[0017] (3)

[0018] In the formula: This represents the largest data processing capacity in the data center.

[0019] Based on satisfying the above constraints Compensation costs for different time periods Represented as:

[0020] (4)

[0021] In the formula: The unit compensation cost for data center demand response load; for Initial task load of the data center at any given moment;

[0022] Energy storage model:

[0023] The constraints that energy storage devices must meet during operation include:

[0024] (5)

[0025] (6)

[0026] (7)

[0027] Equation (5) represents the charging and discharging power constraint for energy storage. and These are the maximum allowable charging and discharging power of energy storage; The value represents the charge / discharge state of energy storage, with 1 indicating charging and 0 indicating discharging; Equation (6) represents the state of charge constraint of energy storage. This represents the initial capacity of the energy storage. and These represent the maximum / minimum remaining capacity allowed for energy storage. and Equation (7) represents the charge / discharge efficiency of the energy storage unit; Equation (7) represents the initial and final state capacity constraints of the energy storage.

[0028] The operating cost of energy storage includes investment cost and operation and maintenance cost, and the average charge and discharge cost is expressed as:

[0029] (8)

[0030] In the formula: This is the converted unit charge / discharge cost;

[0031] Diesel generator model:

[0032] (9)

[0033] (10)

[0034] In the formula: and This is the lower limit of the diesel generator's output. for Output power of the diesel generator during the time period; and The uphill and downhill gradeability of the diesel generator;

[0035] Diesel generators, used as backup power sources for data centers, have generation costs that primarily consist of two parts: operation and maintenance costs and fuel costs.

[0036] (11)

[0037] In the formula: for The output power of the diesel engine at any given time; This is the operating cost coefficient for diesel generators; This refers to the fuel consumption per unit of power generation for a diesel generator. For diesel prices;

[0038] Power grid interaction power model:

[0039] (12)

[0040] In the formula: and They are respectively Power purchased and sold from the power grid during specific time periods; and These represent the upper limits of the power capacity for purchasing and selling electricity; for Electricity purchase and sale status during a given time period;

[0041] exist Interaction costs between virtual power plants and the power grid over time periods Represented as:

[0042] (13)

[0043] In the formula: This refers to the day-ahead trading price of electricity on the power grid;

[0044] Cogeneration unit model:

[0045] A combined heat and power (CHP) unit is a device that simultaneously generates electricity and heat, and its model is represented as:

[0046] (14)

[0047] In the formula: for The natural gas power input to the cogeneration unit during a given period; and These refer to the electrical and thermal conversion efficiencies of the combined heat and power (CHP) unit, respectively. and These refer to the electrical and thermal output power of the combined heat and power unit, respectively. and These are the minimum and maximum input power of the combined heat and power unit, respectively; and These refer to the lower and upper limits of the ramp for combined heat and power units;

[0048] The cost of a combined heat and power (CHP) unit is expressed as follows:

[0049] (15)

[0050] In the formula: This refers to the unit price of natural gas in the natural gas market. The calorific value of natural gas; The power generation efficiency of the CHP unit;

[0051] Gas boiler model:

[0052] A gas-fired boiler supplies energy to the heat load by burning natural gas, and its model can be represented as follows:

[0053] (16)

[0054] In the formula: For GB in Input gas power during the time period; For GB in Output thermal power during the time period; Energy conversion efficiency of GB; and These are the minimum and maximum input air power of GB, respectively; and These represent the lower and upper limits of the ramp rate for GB, respectively.

[0055] The cost of a gas-fired boiler is expressed as follows:

[0056] (17)

[0057] In the formula: for The heating power of the boiler at all times; For the boiler's heating efficiency;

[0058] Wind and solar power generation units:

[0059] (18)

[0060] (19)

[0061] In the formula: and They are respectively The output power of photovoltaic power generation and wind power generation at any given time; and This represents the upper limit of the output power;

[0062] Power balance constraints:

[0063] (20)

[0064] In the formula: for The remaining electrical loads in the virtual power plant during a given time period, excluding data center task loads;

[0065] Thermal equilibrium constraint:

[0066] (twenty one)

[0067] In the formula: for Thermal load of virtual power plant during time period.

[0068] Furthermore, in step S2:

[0069] Green Certificate Trading Mechanism:

[0070] The mathematical model for green certificate transaction costs is shown below:

[0071] (twenty two)

[0072] (twenty three)

[0073] In the formula: For green certificate quota indicators; Green certificates obtained for generating electricity from renewable energy sources; The transaction price per unit of green certificate; The quota coefficient for the demand for green certificates; The conversion factor for converting renewable energy generation into the number of green certificates; , They are respectively Real-time virtual power plant electrical load power and renewable energy output; For green certificate transaction costs;

[0074] Tiered carbon trading mechanism:

[0075] The carbon emissions of virtual power plants mainly come from grid purchases, diesel generator output, and output from CHP and GB units, with carbon emission trading being the primary source. Represented as:

[0076] (twenty four)

[0077] (25)

[0078] In the formula: , These are the carbon emission allowances and actual carbon emissions of the virtual power plant, respectively. , , These are the carbon emission allowances for each unit of electricity supplied by coal-fired power plants, each unit of electricity supplied by diesel-fired power plants, and each unit of heat supplied by natural gas-fired power plants. , , These are the carbon emission coefficients for electricity generated per unit of coal-fired power, electricity generated per unit of diesel-fired power, and heat generated per unit of natural gas-fired power, respectively. The conversion coefficient between electrical energy and heat energy; for The amount of electricity purchased by the virtual power plant at any time.

[0079] Furthermore, the green certificate-step carbon bidirectional interaction model in step S2 is expressed as follows:

[0080] Interactive carbon emission trading volume Represented as:

[0081] (26)

[0082] In the formula: Carbon emission reductions per unit of green certificate; The amount of carbon emission reduction corresponding to the amount of green certificates obtained;

[0083] Tiered carbon trading costs considering the two-way interaction mechanism of green certificates and tiered carbon trading This can be expressed as a piecewise function:

[0084] (27)

[0085] In the formula: It serves as the benchmark price for tiered carbon trading; , These are the compensation coefficient and penalty factor for tiered carbon trading, respectively. The interval length;

[0086] The carbon emission reductions from new energy supply, as reflected in green certificates, can offset part of the carbon emissions in carbon emission rights assessments, thus affecting carbon trading. After adding reward and penalty coefficients, a tiered reward and penalty system for green certificate quotas is obtained. :

[0087] (28)

[0088] In the formula: The interval length; This refers to the green certificate quota reward and penalty coefficient, and has ;

[0089] At this point, the formula for calculating the transaction cost of green certificates is:

[0090] (29)

[0091] Furthermore, the construction of the two-stage robust optimization model in step S3 includes:

[0092] The operational objective of a virtual power plant is to minimize system operating costs.

[0093] (30)

[0094] When the uncertainties of wind and solar power output and load power are not considered, the deterministic model of the virtual power plant optimization scheduling problem is as follows:

[0095] (31)

[0096] In the formula: , To optimize variables:

[0097] (32)

[0098] In equation (31), This is the column vector of coefficients corresponding to the objective function; , , , and These are the coefficient matrices of the variables under the corresponding constraints; , The first row represents the constant column vector; the second row represents the inequality constraints in the optimization model; the third row corresponds to equations (5) and (12); the fourth row represents the predicted values ​​of photovoltaic output and load power in the deterministic optimization model, where:

[0099] (33)

[0100] In the formula: , , They represent Forecast values ​​of photovoltaic power output, wind power output, and load power for the specified time period;

[0101] Robust optimization describes the uncertainty of predicted values ​​by constructing an uncertainty set, which is specifically modeled as follows:

[0102] (34)

[0103] In the formula: , and To account for uncertainties in wind and solar power output and load power; , and These represent the maximum allowable errors for wind and solar power output and load power, respectively.

[0104] The purpose of a two-stage robust optimization model is to find uncertain variables. In the uncertain set The economically optimal scheduling scheme under the worst-case scenario of inward variation is represented by the model as follows:

[0105] (35)

[0106] In the formula: minimizing the outer layer represents the first-stage problem, and the decision variables are... The decision-making process involves power purchase and sales plans and energy storage charging and discharging plans. The inner-layer minimization problem is the second stage problem. After the first stage is completed, the worst-case source-load scenario that maximizes and minimizes operating costs is sought, with the decision variables being... and ; Indicates a given set Time optimization variables The feasible region is expressed as follows:

[0107] (36)

[0108] In the formula: , , , Let represent the dual variables corresponding to each constraint in the minimization problem of the second stage;

[0109] For each set of given uncertain variables Equation (35) can be simplified to the deterministic optimization model shown in Equation (31).

[0110] Furthermore, the process of decomposing the two-stage robust optimization model into a main problem and sub-problems in step S4 includes:

[0111] Decomposing equation (35) yields the main problem in the following form:

[0112] (37)

[0113] In the formula: This represents the current iteration number; For the first Solution of the subproblem after the next iteration; For the first The uncertainty variables obtained after the second iteration under the worst-case scenario The value of ;

[0114] The objective function of the decomposed subproblems is:

[0115] (38)

[0116] Minimizing the inner layer of equation (38) under given (x,u) is a linear optimization problem. According to strong duality theory and the correspondence of equation (36), it can be transformed into a max form, resulting in the following single-layer optimization problem:

[0117] (39)

[0118] The formula contains bilinear terms. The optimal solution to the dual problem corresponds to For an uncertain set One extreme point. When the photovoltaic output reaches the minimum value of the interval and the load power reaches the maximum value of the interval, the operating cost of the virtual power plant is higher, which is more in line with the definition of the "worst-case scenario". Therefore, equation (34) can be rewritten as follows:

[0119] (40)

[0120] In the formula: For binary variables, a value of 1 indicates that the uncertain variable for the corresponding time period reaches the boundary of the interval; , and These are the uncertainty adjustment parameters for photovoltaic power output, wind power output, and load power, respectively, with values ​​ranging from... The integers within represent the total number of time periods during which the photovoltaic output and load power reach the boundary of the deviation interval described by equation (40) within the scheduling cycle. Substituting the expression for the uncertain variable in equation (40) into equation (39) will result in a product of binary and continuous variables. By introducing auxiliary variables and relevant constraints to linearize it, we can obtain:

[0121] (41)

[0122] In the formula: ; For the introduction of continuous auxiliary variables; The upper bound of the dual variable can be a sufficiently large positive real number.

[0123] This invention guides data centers to participate in demand response through a price compensation mechanism, and incorporates the diesel generators equipped in data centers as supplementary power sources during peak electricity demand periods when grid power purchases are difficult, thus establishing a green certificate-tiered carbon bidirectional interaction model. Data center demand response load refers to the data center's strategy of incentivizing users to provide deferred batch processing tasks through reasonable price subsidies, thereby optimizing load scheduling in terms of time. Diesel generators provide supplementary power to meet the load demand of the virtual power plant during peak electricity demand periods when grid power purchases are difficult.

[0124] The method of this invention optimizes the scheduling of a virtual power plant that includes data center demand response load and standby diesel generators through a two-stage robust optimization model. It uses uncertainty adjustment parameters to flexibly adjust the conservatism of the scheduling scheme, and further introduces a green certificate-tiered carbon trading mechanism to obtain an optimized scheduling scheme with the goal of minimizing the overall operating cost of the virtual power plant.

[0125] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0126] (1) This invention encourages users to provide some of the tasks that can be delayed by price compensation, thereby guiding the data center load to participate in demand response and effectively improving the data center load adjustment capability.

[0127] (2) This invention aggregates the data center load and its backup power sources such as energy storage and diesel generators into a virtual power plant, and establishes a two-stage robust model with the goal of minimizing costs, thereby improving the economic efficiency of data center operation while taking into account uncertainties.

[0128] (3) Based on the carbon emission reduction effect of green certificates, this invention establishes a two-way interaction model of green certificates and tiered carbon, and further introduces it into a virtual power plant, which effectively improves the renewable energy absorption capacity. Attached Figure Description

[0129] Figure 1 This is a flowchart of the method of the present invention;

[0130] Figure 2 This is a schematic diagram of the system structure of the virtual power plant of the present invention;

[0131] Figure 3 It is the curve of predicted electrical load and the curve of actual electrical load inside the virtual power plant;

[0132] Figure 4These are the predicted and actual values ​​for wind and solar power generation:

[0133] Figure 5 This is a diagram showing the results of optimized electrical load scheduling in a typical scenario;

[0134] Figure 6 This is a diagram showing the results of hot load optimization scheduling in a typical scenario. Detailed Implementation

[0135] The present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. After reading this invention, any modifications of the invention in various equivalent forms by those skilled in the art will fall within the scope defined by the appended claims.

[0136] Example 1:

[0137] like Figure 1 As shown, this embodiment provides a control method for a virtual power plant with diesel generation based on green certificate-tiered carbon bidirectional interaction, including the following steps:

[0138] S1: Construct a virtual power plant model by comprehensively considering the physical characteristics of each device in the virtual power plant;

[0139] S2: Introduce green certificate trading mechanism and tiered carbon trading mechanism to construct a two-way interaction model of green certificate-tiered carbon;

[0140] S3: Based on the virtual power plant model and the green certificate-tiered carbon bidirectional interaction model, an optimization model is established with the goal of minimizing the overall operating cost of the virtual power plant plus the transaction cost of green certificates-tiered carbon, and it is rewritten as a two-stage robust optimization model with a min-max-min structure.

[0141] S4: Using the column constraint generation algorithm and strong duality theory, the two-stage robust optimization model is decomposed into a main problem and sub-problems for iterative solution, thereby obtaining a virtual power plant optimization scheduling scheme that considers low-carbon benefits.

[0142] The virtual power plant model in step S1 includes a data center load model, an energy storage model, a diesel generator model, a grid interaction power model, a combined heat and power unit model, a gas boiler model, wind and photovoltaic power generation units, and power balance constraints and heat balance constraints, as specifically expressed below:

[0143] Data center load model:

[0144] Data center tasks are typically categorized into two types: real-time tasks and batch tasks. Real-time tasks require immediate processing, while batch tasks allow for delayed processing within a certain timeframe. By participating in demand response, data centers can dynamically adjust batch task loads, reducing energy consumption during peak hours and maximizing power availability during off-peak hours. By offering price compensation, data centers can receive a higher proportion of batch tasks. The power model for a data center can be expressed as follows:

[0145] (1)

[0146] (2)

[0147] In the formula: This is a variable representing the proportion of the total task load where users convert real-time tasks to batch processing tasks due to price compensation. The number of batch processing tasks. The first report submitted by the user The maximum delay time for each task for The batch processing task load received by the data center at any given moment; for Real-time task load received by the data center at any given moment. and Assigned to The sum of the task load at each moment is the current task load. Total data center workload at any given time ;

[0148] The workload in a data center should be less than the data center's maximum processing capacity.

[0149] (3)

[0150] In the formula: This represents the largest data processing capacity in the data center.

[0151] Based on satisfying the above constraints Compensation costs for different time periods Represented as:

[0152] (4)

[0153] In the formula: The unit compensation cost for data center demand response load; for Initial task load of the data center at any given moment;

[0154] Energy storage model:

[0155] The constraints that energy storage devices must meet during operation include:

[0156] (5)

[0157] (6)

[0158] (7)

[0159] Equation (5) represents the charging and discharging power constraint for energy storage. and These are the maximum allowable charging and discharging power of energy storage; The value represents the charge / discharge state of energy storage, with 1 indicating charging and 0 indicating discharging; Equation (6) represents the state of charge constraint of energy storage. This represents the initial capacity of the energy storage. and These represent the maximum / minimum remaining capacity allowed for energy storage. and Equation (7) represents the charge / discharge efficiency of the energy storage unit; Equation (7) represents the initial and final state capacity constraints of the energy storage.

[0160] The operating cost of energy storage includes investment cost and operation and maintenance cost, and the average charge and discharge cost is expressed as:

[0161] (8)

[0162] In the formula: This is the converted unit charge / discharge cost;

[0163] Diesel generator model:

[0164] Diesel generators can serve as a controllable power source to provide electricity to virtual power plants when other energy sources cannot meet load demands.

[0165] (9)

[0166] (10)

[0167] In the formula: and This is the lower limit of the diesel generator's output. for Output power of the diesel generator during the time period; and The uphill and downhill gradeability of the diesel generator;

[0168] Diesel generators, used as backup power sources for data centers, have generation costs that primarily consist of two parts: operation and maintenance costs and fuel costs.

[0169] (11)

[0170] In the formula: for The output power of the diesel engine at any given time; This is the operating cost coefficient for diesel generators; This refers to the fuel consumption per unit of power generation for a diesel generator. For diesel prices;

[0171] Power grid interaction power model:

[0172] (12)

[0173] In the formula: and They are respectively Power purchased and sold from the power grid during specific time periods; and These represent the upper limits of the power capacity for purchasing and selling electricity; for Electricity purchase and sale status during a given time period;

[0174] exist Interaction costs between virtual power plants and the power grid over time periods Represented as:

[0175] (13)

[0176] In the formula: This refers to the day-ahead trading price of electricity on the power grid;

[0177] Combined Heat and Power Unit Model (CHP):

[0178] A combined heat and power (CHP) unit is a device that simultaneously generates electricity and heat, and its model is represented as:

[0179] (14)

[0180] In the formula: for The natural gas power input to the cogeneration unit during a given period; and These refer to the electrical and thermal conversion efficiencies of the combined heat and power (CHP) unit, respectively. and These refer to the electrical and thermal output power of the combined heat and power unit, respectively. and These are the minimum and maximum input power of the combined heat and power unit, respectively; and These refer to the lower and upper limits of the ramp for combined heat and power units;

[0181] The cost of a combined heat and power (CHP) unit is expressed as follows:

[0182] (15)

[0183] In the formula: This refers to the unit price of natural gas in the natural gas market. The calorific value of natural gas; The power generation efficiency of the CHP unit;

[0184] Gas boiler model:

[0185] A gas-fired boiler supplies energy to the heat load by burning natural gas, and its model can be represented as follows:

[0186] (16)

[0187] In the formula: For GB in Input gas power during the time period; For GB in Output thermal power during the time period; Energy conversion efficiency of GB; and These are the minimum and maximum input air power of GB, respectively; and These represent the lower and upper limits of the ramp rate for GB, respectively.

[0188] The cost of a gas-fired boiler is expressed as follows:

[0189] (17)

[0190] In the formula: for The heating power of the boiler at all times; For the boiler's heating efficiency;

[0191] Wind and solar power generation units:

[0192] (18)

[0193] (19)

[0194] In the formula: and They are respectively The output power of photovoltaic power generation and wind power generation at any given time; and This represents the upper limit of the output power;

[0195] Power balance constraints:

[0196] (20)

[0197] In the formula: for The remaining electrical loads in the virtual power plant during a given time period, excluding data center task loads;

[0198] Thermal equilibrium constraint:

[0199] (twenty one)

[0200] In the formula: for Thermal load of virtual power plant during time period.

[0201] The green certificate-tiered carbon trading mechanism in step S2 is represented as follows:

[0202] Green Certificate Trading Mechanism (GCT):

[0203] The mathematical model for green certificate transaction costs is shown below:

[0204] (twenty two)

[0205] (twenty three)

[0206] In the formula: For green certificate quota indicators; Green certificates obtained for generating electricity from renewable energy sources; The transaction price per unit of green certificate; The quota coefficient for the demand for green certificates; The conversion factor for converting renewable energy generation into the number of green certificates; , They are respectively Real-time virtual power plant electrical load power and renewable energy output; For green certificate transaction costs;

[0207] Tiered carbon trading mechanism (CET):

[0208] The carbon emissions of virtual power plants mainly come from grid purchases, diesel generator output, and output from CHP and GB units, with carbon emission trading being the primary source. Represented as:

[0209] (twenty four)

[0210] (25)

[0211] In the formula: , These are the carbon emission allowances and actual carbon emissions of the virtual power plant, respectively. , , These are the carbon emission allowances for each unit of electricity supplied by coal-fired power plants, each unit of electricity supplied by diesel-fired power plants, and each unit of heat supplied by natural gas-fired power plants. , , These are the carbon emission coefficients for electricity generated per unit of coal-fired power, electricity generated per unit of diesel-fired power, and heat generated per unit of natural gas-fired power, respectively. The conversion coefficient between electrical energy and heat energy; for The amount of electricity purchased by the virtual power plant at any time.

[0212] The green certificate-step carbon bidirectional interaction model is expressed as follows:

[0213] Interactive carbon emission trading volume Represented as:

[0214] (26)

[0215] In the formula: Carbon emission reductions per unit of green certificate; The amount of carbon emission reduction corresponding to the amount of green certificates obtained;

[0216] Tiered carbon trading costs considering the two-way interaction mechanism of green certificates and tiered carbon trading This can be expressed as a piecewise function:

[0217] (27)

[0218] In the formula: It serves as the benchmark price for tiered carbon trading; , These are the compensation coefficient and penalty factor for tiered carbon trading, respectively. The interval length;

[0219] The carbon emission reductions from new energy supply, as reflected in green certificates, can offset part of the carbon emissions in carbon emission rights assessments, thus affecting carbon trading. After adding reward and penalty coefficients, a tiered reward and penalty system for green certificate quotas is obtained. :

[0220] (28)

[0221] In the formula: The interval length; This refers to the green certificate quota reward and penalty coefficient, and has ;

[0222] At this point, the formula for calculating the transaction cost of green certificates is:

[0223] (29)

[0224] The construction of the two-stage robust optimization model in step S3 includes:

[0225] The operational objective of a virtual power plant is to minimize system operating costs.

[0226] (30)

[0227] When the uncertainties of wind and solar power output and load power are not considered, the deterministic model of the virtual power plant optimization scheduling problem is as follows:

[0228] (31)

[0229] In the formula: , To optimize variables:

[0230] (32)

[0231] In equation (31), This is the column vector of coefficients corresponding to the objective function; , , , and These are the coefficient matrices of the variables under the corresponding constraints; , The first row represents the constant column vector; the second row represents the inequality constraints in the optimization model; the third row corresponds to equations (5) and (12); the fourth row represents the predicted values ​​of photovoltaic output and load power in the deterministic optimization model, where:

[0232] (33)

[0233] In the formula: , , They represent Forecast values ​​of photovoltaic power output, wind power output, and load power for the specified time period;

[0234] Robust optimization describes the uncertainty of predicted values ​​by constructing an uncertainty set, which is specifically modeled as follows:

[0235] (34)

[0236] In the formula: , and To account for uncertainties in wind and solar power output and load power; , and These represent the maximum allowable errors for wind and solar power output and load power, respectively.

[0237] The purpose of a two-stage robust optimization model is to find uncertain variables. In the uncertain set The economically optimal scheduling scheme under the worst-case scenario of inward variation is represented by the model as follows:

[0238] (35)

[0239] In the formula: minimizing the outer layer represents the first-stage problem, and the decision variables are... The decision-making process involves power purchase and sales plans and energy storage charging and discharging plans. The inner-layer minimization problem is the second stage problem. After the first stage is completed, the worst-case source-load scenario that maximizes and minimizes operating costs is sought, with the decision variables being... and ; Indicates a given set Time optimization variables The feasible region is expressed as follows:

[0240] (36)

[0241] In the formula: , , , Let represent the dual variables corresponding to each constraint in the minimization problem of the second stage;

[0242] For each set of given uncertain variables Equation (35) can be simplified to the deterministic optimization model shown in Equation (31).

[0243] In step S4, the column constraint generation algorithm (C&CG) is used to solve the two-stage robust optimization model described above. The C&CG algorithm decomposes the original problem into a main problem and sub-problems, which are solved alternately until the objective function value reaches the convergence threshold, thus obtaining the final result of the two-stage robust optimization model.

[0244] The process of breaking down a two-stage robust optimization model into a main problem and sub-problems includes:

[0245] Decomposing equation (35) yields the main problem in the following form:

[0246] (37)

[0247] In the formula: This represents the current iteration number; For the first Solution of the subproblem after the next iteration; For the first The uncertainty variables obtained after the second iteration under the worst-case scenario The value of ;

[0248] The objective function of the decomposed subproblems is:

[0249] (38)

[0250] Minimizing the inner layer of equation (38) under given (x,u) is a linear optimization problem. According to strong duality theory and the correspondence of equation (36), it can be transformed into a max form, resulting in the following single-layer optimization problem:

[0251] (39)

[0252] The formula contains bilinear terms. The optimal solution to the dual problem corresponds to For an uncertain set One extreme point. When the photovoltaic output reaches the minimum value of the interval and the load power reaches the maximum value of the interval, the operating cost of the virtual power plant is higher, which is more in line with the definition of the "worst-case scenario". Therefore, equation (34) can be rewritten as follows:

[0253] (40)

[0254] In the formula: For binary variables, a value of 1 indicates that the uncertain variable for the corresponding time period reaches the boundary of the interval; , and These are the uncertainty adjustment parameters for photovoltaic power output, wind power output, and load power, respectively, with values ​​ranging from... The integers within represent the total number of time periods during which the photovoltaic output and load power reach the boundary of the deviation interval described by equation (40) within the scheduling cycle. Substituting the expression for the uncertain variable in equation (40) into equation (39) will result in a product of binary and continuous variables. By introducing auxiliary variables and relevant constraints to linearize it, we can obtain:

[0255] (41)

[0256] In the formula: ; For the introduction of continuous auxiliary variables; The upper bound of the dual variable can be a sufficiently large positive real number.

[0257] The C&CG algorithm is used to solve the main problem and its subproblems. The process is as follows:

[0258] 1) Given a set of uncertain variables as the initial worst-case scenario;

[0259] 2) Set the lower bound Upper Realm The number of iterations K=1;

[0260] 3) Solve the main problem based on the worst-case scenario to obtain the optimal solution;

[0261] 4) Use the objective function value obtained from the main problem as the new upper bound U;

[0262] 5) Substitute the solution obtained from the main problem into the subproblem and solve the subproblem;

[0263] 6) Obtain the objective function value of the subproblem and update the lower bound L;

[0264] 7) Judgment If the convergence threshold has been reached, output the optimal solution; otherwise, add auxiliary variables and related constraints, K=K+1, and return to step 3.

[0265] Example 2:

[0266] This embodiment uses Figure 2 The virtual power plant shown is used as a case study to verify the effectiveness of the model proposed in this invention, as detailed below:

[0267] The optimized scheduling cycle is 24 hours, with a time step of 1 hour; time-of-use pricing is shown in Table 1, equipment parameters are shown in Table 2, and other parameters are shown in Table 3. The virtual power plant load prediction curve and actual values ​​are shown below. Figure 3 As shown, the predicted curves for wind and solar power generation and the actual values ​​are as follows: Figure 4 As shown, the load power uncertainty adjustment parameter is set to 12, the photovoltaic uncertainty adjustment parameter is set to 6, and the wind power uncertainty adjustment parameter is set to 12. The uncertainty prediction deviations for load power and wind and solar power output are 10% and 15% of the predicted values, respectively.

[0268] Table 1 Time-of-use electricity prices

[0269]

[0270] Table 2 Equipment Parameters

[0271]

[0272] Table 3 Other parameters

[0273]

[0274] To verify the effectiveness of the proposed green certificate-tiered carbon bidirectional interaction mechanism, four scenarios were set up: Scenario 1, without considering green certificate trading and tiered carbon trading; Scenario 2, considering tiered carbon trading based on Scenario 1; Scenario 3, comprehensively considering tiered carbon trading and green certificate trading based on Scenario 1; Scenario 4, considering the proposed green certificate-tiered carbon bidirectional interaction mechanism. The optimization results of the virtual power plant in the four scenarios are shown in Table 4. The comprehensive operating costs of the virtual power plant in each scenario were compared to verify the effectiveness of the proposed green certificate-tiered carbon bidirectional interaction mechanism. In the table: negative values ​​for tiered carbon cost and green certificate cost represent tiered carbon revenue and green certificate revenue, respectively. As shown in Table 4, Scenario 4 has the lowest comprehensive operating cost, which is reduced by RMB 3950.82, RMB 4457.62, and RMB 3782.11 compared to Scenarios 1, 2, and 3, respectively. The above results demonstrate that adopting the green certificate-tiered carbon two-way interaction mechanism can improve system economics, reduce carbon emissions, and promote the low-carbon economic operation of virtual power plants.

[0275] Table 4 Costs in Different Scenarios

[0276]

[0277] Building upon Scenario 4, this paper further analyzes the impact of robust optimization algorithms and demand response on the overall operating cost of the virtual power plant. Three scenarios are set up: Scenario 5, using deterministic optimization without considering demand response; Scenario 6, using robust optimization without considering demand response; and Scenario 7, using robust optimization and considering demand response. The optimization results of the virtual power plant in the three scenarios are shown in Table 5. Table 5 shows that the overall operating cost of the scheduling scheme obtained using deterministic optimization is lower than that obtained using robust optimization. This is because robust optimization considers the impact of uncertainty. Since the real-time market electricity purchase / sale price is generally higher / lower than the day-ahead market price, ignoring the impact of uncertainty would require the virtual power plant to compensate for prediction errors in the real-time market, thus increasing the final overall operating cost of the virtual power plant. Comparing Scenario 6 and Scenario 7, it can be seen that the overall operating cost of the virtual power plant is significantly reduced after considering data center demand response. This is because the data center postpones the processing time of batch processing tasks, allowing the virtual power plant to minimize electricity demand during peak electricity price periods, thereby reducing the overall operating cost of the virtual power plant.

[0278] Table 5 Cost Analysis for Different Scenarios

[0279]

[0280] Therefore, by adopting the optimized scheduling method of this invention, combined with the data center demand response and green certificate-tiered carbon bidirectional interaction mechanism, the operating cost of virtual power plants can be effectively reduced, the robustness of the scheduling scheme and its ability to resist the risk of real-time market electricity price fluctuations can be improved, and significant low-carbon economic benefits can be achieved, based on the uncertainty of virtual power plant source and load.

[0281] The optimized scheduling scheme for scenario 4 is as follows: Figure 5 As shown, from 1:00 to 5:00, electricity prices are low. During this period, the virtual power plant mainly meets the system's load demand through wind power generation, CHP unit generation, and purchasing electricity from the grid, while energy storage is charging. From 6:00 to 9:00, photovoltaic output gradually increases, and energy storage devices enhance the absorption of wind and solar power output through their own charging and discharging characteristics. From 10:00 to 14:00, due to the high pressure on the grid supply during peak electricity demand, the virtual power plant's purchase of electricity from the grid is limited. At this time, it needs to supplement its output through diesel generator generation and energy storage discharge. From 15:00 to 17:00... At 0:00, electricity prices are at parity, and energy storage is recharged. From 18:00 to 20:00, electricity demand peaks again, requiring diesel generators to generate electricity and energy storage to discharge to supplement output. Due to the ramp-up constraints of diesel generators, they need to ramp up to a higher output level before peak demand to meet peak load demand. From 21:00 to 24:00, wind power and CHP units generate electricity, and electricity is purchased from the grid to meet system load demand. Meanwhile, energy storage needs to be charged during this period to meet capacity constraints at both the beginning and end of the cycle.

[0282] The heat load demand is mainly met by combined heat and power units, with gas-fired boilers providing supplementary output, resulting in a heat-power balance. Figure 6 As shown.

Claims

1. A control method for a virtual power plant containing biomass based on green certificate-ladder carbon two-way interaction, characterized in that, The method comprises the following steps: S1: comprehensively considering the physical characteristics of each device in the virtual power plant, constructing a virtual power plant model; S2: introducing a two-way interaction mechanism, constructing a new green certificate-ladder carbon two-way interaction model; S3: based on the virtual power plant model and the two-way interaction model, establishing an optimization model with the minimum overall operation cost of the virtual power plant and the minimum transaction cost as the target, and rewriting it into a robust optimization model with a min-max-min structure; S4: using column constraint generation algorithm and strong duality theory to split the robust optimization model into a master problem and a sub-problem; using C&CG algorithm to solve the master problem and the sub-problem, and obtaining a virtual power plant control scheme.

2. The control method of the virtual power plant containing the biomass based on the green certificate-ladder carbon two-way interaction according to claim 1, characterized in that, In the virtual power plant model of the step S1: The power model of the data center is expressed as: The task load of the data center should be less than the maximum processing capacity of the data center: （1）； （2）； In the formula: is the proportion of the total task load that is converted from real-time tasks to batch tasks due to the influence of user price compensation, is the number of batch tasks, is the maximum delay time of the first task reported by the user, is the batch task load received by the data center at time t; is the real-time task load received by the data center at time t, and the sum of the task loads allocated to time t is the total task load of the data center at the current time t ; The operation cost of the energy storage includes investment cost and operation and maintenance cost, and the average charging and discharging cost is expressed as: （3）； In the formula: is the maximum data processing capacity of the data center; On the basis of meeting the above constraints, Compensation costs for time periods Is expressed as: （4）； In the formula: is the unit compensation cost of the data center demand response load; is is the initial task load of the data center at the moment The diesel generator model: The diesel generator is a backup power source equipped for the data center, and the power generation cost considers two parts of operation and maintenance cost and fuel cost, that is: （5）； （6）； （7）； is the charge-discharge power constraint of the energy storage, and are the maximum charge-discharge power allowed by the energy storage, respectively; is the charge-discharge state of the energy storage, which takes the value of 1 when charging and 0 when discharging; equation (6) represents the state of charge constraint of the energy storage, is the initial capacity of the energy storage, and are the maximum / minimum residual capacity allowed by the energy storage, respectively, and are the charge-discharge efficiencies of the energy storage unit; equation (7) is the initial-final state capacity constraint of the energy storage; In the virtual power plant model of the step S1: （8）； In the formula: is the unit cost of charging and discharging after conversion; The grid interaction power model: （9）； （10）； wherein: and is the lower output limit of the diesel generator; is is the output power of the diesel generator for a period of time; and is the ramp-up and ramp-down rate of the diesel generator; The combined heat and power unit model: （11）； In the formula: is the output power of the diesel engine at the moment; is the diesel generator operation cost coefficient; is the unit power generation oil consumption of the diesel generator; is the diesel price.

3. The control method of the virtual power plant containing the diesel according to claim 2, wherein, The combined heat and power unit is a device that simultaneously generates electric energy and heat energy, and its model is expressed as: The cost of the combined heat and power unit is expressed as: （12）； wherein: and respectively represent the upper limit of the power of the purchased and sold electricity; the period of time from the grid for the purchased and sold electricity power; and respectively represent the upper limit of the power of the purchased and sold electricity; is the state of the purchased and sold electricity for the period of time; In Interacting cost of virtual power plant and grid over time period is represented as: （13）； In the formula: is the day-ahead transaction price of the power grid; In the virtual power plant model of the step S1: The gas boiler model: （14）； wherein: is the power input of the CCHP unit; is the power input of the CCHP unit; and are the electrical and thermal conversion efficiencies of the CCHP unit, respectively; and are the electrical and thermal output powers of the CCHP unit, respectively; and are the minimum and maximum input powers of the CCHP unit, respectively; and are the lower and upper limits of the ramp of the CCHP unit, respectively. The gas boiler supplies energy for the heat load by burning natural gas, and its model is expressed as: （15）； In the formula: is the unit price of natural gas in the market; is the heat value of natural gas; is the power generation efficiency of the CHP unit.

4. The control method of the virtual power plant containing the biomass based on the green certificate-ladder carbon two-way interaction according to claim 3, characterized in that, The cost of the gas boiler is expressed as: Wind and photovoltaic power generation units: In the virtual power plant model of the step S1: （16） In the formula: For GB in Input gas power during the time period; For GB in Output thermal power during the time period; Energy conversion efficiency of GB; and These are the minimum and maximum input air power of GB, respectively; and These represent the lower and upper limits of the ramp rate for GB, respectively. The electric energy balance constraint: （17）； In the formula: is the heating power of the boiler at the moment; is the heating efficiency of the boiler; The heat balance constraint: （18）； （19）； In the formula: is the output power of the photovoltaic power generation at the time t, and is the output power of the wind power generation at the time t, respectively. is the output power of the photovoltaic power generation at the time t, and is the output power of the wind power generation at the time t, respectively. is the upper limit of the output power.

5. The control method of the virtual power plant containing the diesel according to claim 4, wherein, The green certificate transaction mechanism in the step S2: The mathematical model of the green certificate transaction cost is as follows: （20）； In the formula: is Period virtual power plant, in addition to the remaining electrical load of data center task load; The ladder carbon transaction mechanism in the step S2: （21）； In the formula: is a heat load of the period virtual power plant.

6. The control method of the virtual power plant containing the biomass based on the green certificate-ladder carbon two-way interaction according to claim 5, characterized in that, The expression of the green certificate-ladder carbon two-way interaction model in the step S2 is: At this time, the calculation formula of the green certificate transaction cost is: （22）； （23）； In the formula, is the green certificate quota index; is the green certificate obtained by renewable energy power generation; is the transaction price of a unit green certificate; is the quota coefficient of green certificate demand; is the conversion coefficient of renewable energy power generation to the number of green certificates; , are respectively is the electric load power of the virtual power plant at the moment, and renewable energy output; is the green certificate transaction cost.

7. The control method of the virtual power plant containing the diesel according to claim 6, wherein, The construction of the two-stage robust optimization model in the step S3 comprises: Carbon emission rights trading is represented as: （24）； （25）； In the formula: , These are the carbon emission allowances and actual carbon emissions of the virtual power plant, respectively. , , These are the carbon emission allowances for each unit of electricity supplied by coal-fired power plants, each unit of electricity supplied by diesel-fired power plants, and each unit of heat supplied by natural gas-fired power plants. , , These are the carbon emission coefficients for electricity generated per unit of coal-fired power, electricity generated per unit of diesel-fired power, and heat generated per unit of natural gas-fired power, respectively. The conversion coefficient between electrical energy and heat energy; for The amount of electricity purchased by the virtual power plant at any time.

8. The control method of the virtual power plant containing the diesel according to claim 7, wherein, The operation target of the virtual power plant is to minimize the system operation cost: Carbon emission right trade volume after interaction is represented as: （26）； In the formula: is the carbon reduction amount of the unit green certificate; is the carbon reduction amount corresponding to the green certificate acquisition amount; Considering the green certificate-ladder carbon two-way interaction mechanism of ladder carbon trading cost In the form of piecewise function is expressed as: （27）； In the formula: is the benchmark price of the tiered carbon trade; , are the compensation factor and the penalty factor of the tiered carbon trade, respectively; is the interval length; The ladder green certificate quota reward and punishment amount is obtained after superimposing the reward and punishment coefficient : （28）； In the formula: is the interval length; is the green certificate quota reward and punishment coefficient, and has ; When the influence of wind and light output and load power uncertainty is not considered, the deterministic model of the virtual power plant optimization scheduling problem is: （29）。 9. The control method of the virtual power plant containing the diesel according to claim 8, wherein, The robust optimization describes the uncertainty of the predicted value by constructing an uncertainty set, and the uncertainty set is modeled as follows: The process of splitting the two-stage robust optimization model into a master problem and a sub-problem in the step S4 comprises: （30）； The master problem obtained by decomposing formula (35) is in the form of: （31）； In the formulae: , are optimization variables: （32）； In formula (31), is a coefficient column vector corresponding to the objective function; , , , and are coefficient matrices corresponding to variables under constraints, respectively; , are constant column vectors; the first row of the constraint condition represents an inequality constraint in the optimization model; the second row is an equality constraint; the third row corresponds to formula (5) and formula (12); the fourth row indicates that in the deterministic optimization model, the values of the photovoltaic output and the load power are predicted values, wherein: （33）； in which: , , respectively represent forecast values of the photovoltaic power output, the wind power output and the load power for the time period. The objective function of the decomposed sub-problem is: （34）； wherein: , and are the wind and solar power outputs and the load power uncertainty variables considering the uncertainty; , and are the maximum error allowed for the wind and solar power outputs and the load power, respectively. The purpose of the two-stage robust optimization model is to find the economic optimal dispatching scheme under the worst scenario change of the uncertain variables in the uncertain set The model is expressed as: （35）； where the minimization of the outer layer is the first stage problem with decision variables , and the decision of the electricity purchase and sale plan and the energy storage charging and discharging plan; the maximization of the inner layer is the second stage problem, after the first stage is completed, the worst source and load scenario is found to maximize the minimum value of the operation cost, and the decision variable is and ; represents the feasible region of the optimization variable when a given set of is given, and the specific expression is as follows: （36）； where: , , , denote the dual variables corresponding to each constraint in the minimization problem of the second stage. For each given set of uncertain variables , equation (35) can be reduced to the deterministic optimization model shown in equation (31).

10. The control method of the virtual power plant containing the diesel according to claim 9, wherein, The inner minimization of formula (38) under given (x, u) is a linear optimization problem, which can be converted into max form according to the strong duality theory and the corresponding relationship of formula (36), and the following single-layer optimization problem is obtained: ​ （37）； wherein: is the current iteration number; is the solution of the subproblem after the iteration; is the value of the uncertain variable after the iteration in the worst-case scenario; and is the value of the uncertain variable ​ （38）； ​ （39）； where the bilinear term is present The optimal solution of the dual problem corresponds to is an extreme point of the uncertainty set When the photovoltaic output takes the minimum value of the interval and the load power takes the maximum value of the interval, the operation cost of the virtual power plant is higher, which is more in line with the definition of "worst-case scenario", so formula (34) can be rewritten as follows: （40）； wherein: is a binary variable, taking value 1 when the uncertain variable of the corresponding time period takes the boundary of the interval; , and are the uncertainty adjustment parameters of photovoltaic output, wind power output and load power, respectively, taking integer values in the range of , representing the total number of time periods in the scheduling period when the photovoltaic output and load power take the boundary of the deviation interval described by equation (40); after substituting the uncertain variable expression in equation (40) into equation (39), the form of binary variable and continuous variable product will appear, which is linearized by introducing auxiliary variables and related constraints to obtain: （41）； where: ; is an introduced continuous auxiliary variable; is an upper bound on the dual variable.

Images