A method for energy storage aggregator participating in day-ahead-real-time power market bidding

Abstract

The application discloses a kind of energy storage polymerization business participation day-to-day real-time power market bidding method, comprising: completing upper model modeling;Complete lower model modeling;Rewrite the objective function and constraint condition in the original problem of lower model into Lagrange dual form, convert double-layer model into single-layer model;Linearization is carried out to the multiplication of dual variable and continuous variable in complementary constraint condition using large M method, and binary expansion method is used to process the multiplication variable of price and charge-discharge capacity in the objective function of upper model;Linearized single-layer model is solved using commercial solver Gurobi, and the optimal bidding strategy parameter and maximum market income are obtained, and the market price parameter and charge-discharge capacity parameter generated therefrom.The application can obtain globally optimal results while ensuring solution speed, solve the problem that energy storage aggregator cannot effectively regulate day-to-day real-time power market, lack flexibility in market competition and have low income.

Description

Technical Field

[0001] This invention belongs to the field of electricity market technology, specifically relating to a method for energy storage aggregators to participate in day-ahead-real-time electricity market bidding. Background Technology

[0002] With the increasing generation of intermittent renewable energy sources such as solar and wind power, the stability of the power system will face increasingly severe challenges in the future. Power operators urgently need a means to effectively reduce system fluctuation risks and avoid the harm caused by generation uncertainty. Thanks to technological advancements in chemistry, materials science, and chemical engineering, energy storage technology has developed rapidly in recent years, with various sizes and types of energy storage devices being built and put into use across the country. Energy storage, due to its safe and adjustable characteristics, can make a significant contribution to power system stability. However, there are currently two difficulties in enabling energy storage devices to participate in power system regulation through market mechanisms: First, distributed energy storage devices are complex, with insufficient individual capacity and wide geographical distribution, making effective control and market competitiveness difficult; second, fluctuations in electricity prices and load demand bring significant uncertainty to the market bidding process, and there is currently a lack of effective bidding strategies suitable for energy storage devices, thus hindering the expansion of settlement revenue and the stimulation of market participation.

[0003] The aggregation of multiple energy storage entities into energy storage aggregators, replacing a wide range of small-capacity energy storage devices in the electricity market, is an important means of addressing the distributed nature of energy storage. Aggregators act as intermediaries, aggregating and managing energy storage devices, participating in market bidding processes, and sharing profits. This simplifies market processes and lowers market entry barriers, and has attracted increasing attention and research.

[0004] Currently, energy storage aggregators face technical challenges in participating in the day-ahead-real-time electricity market, including difficulties in effectively regulating market competition, lack of flexibility, and low returns. Summary of the Invention

[0005] Technical problem solved: This invention addresses the problems of ineffective regulation, lack of market competition flexibility, and low returns for energy storage aggregators participating in the day-ahead-real-time electricity market. It proposes a bidding method for energy storage aggregators to participate in the day-ahead-real-time electricity market based on two-level stochastic programming.

[0006] Technical solution:

[0007] A method for energy storage aggregators to participate in day-ahead-real-time electricity market bidding, the electricity market bidding method comprising the following steps:

[0008] Step 1: The energy storage aggregator obtains cost information and derives the unit power charge / discharge cost function C. bat Energy storage aggregators simultaneously provide electrical energy and reserves to participate in the day-ahead and real-time electricity markets, respectively, and will formulate bidding strategy parameters k. D-1and k D The data is passed down to the lower level, given the range of pricing strategy parameters, to obtain the lower-level price parameter (λ). t ,λ ωt and charge / discharge parameters Based on this, the upper-level model is modeled with the goal of maximizing the market participation benefit W.

[0009] Step 2, obtain the pricing strategy parameters (k) of the energy storage aggregator. D-1 ,k D ), where k D-1 and k D Based on the day-ahead and real-time market pricing strategy parameters, a total generation cost equation is formulated. System power flow constraints are handled using stochastic programming, taking into account the generation and reserve output constraints of energy storage aggregators and other generators within the system. This completes the lower-level model modeling, whose output includes the day-ahead market clearing price λ. t and real-time market clearing price λ ωt Planned charging volume of energy storage aggregators and discharge quantity Energy storage aggregators reduce reserve requirements and reduce reserve And return the results to the upper-level model;

[0010] Step 3: Rewrite the objective function and constraints in the original problem of the lower-level model into Lagrange dual form, and use KKT conditions to replace the lower-level model with the original problem feasibility conditions, the dual problem feasibility conditions, and complementary relaxation constraints, thus transforming the two-level model into a single-level model. In the transformed single-level model, the objective function is the same as the original upper-level model objective function, and the constraints include the original upper-level model constraints, the original problem feasibility conditions of the lower-level model, the dual problem feasibility conditions of the lower-level model, and the complementary relaxation constraints of the lower-level problem.

[0011] Step 4: The Big M method is used to linearize the multiplication of dual and continuous variables in the complementary constraints. The binary expansion method is used to linearize the multiplication of price and charge / discharge quantities in the objective function of the upper-level model. and Process it;

[0012] Step 5: Solve the linearized single-layer model using the commercial solver Gurobi to obtain the optimal pricing strategy parameters. and maximum market return W max and the resulting market price parameters and charge / discharge parameters

[0013] Furthermore, step 1 specifically includes the following sub-steps:

[0014] Step 1.1, the energy storage aggregator obtains its unit power charge / discharge cost function C. bat According to its pricing strategy parameter (k) D-1 ,k D ), respectively according to k D-1 ·C bat and k D ·C bat Price quotes are provided for two usage scenarios: electrical energy and backup power, given the known lower-level price parameter (λ). t ,λ ωt and charge / discharge parameters Based on this, its market participation return W is expressed as follows:

[0015]

[0016] In the formula, t represents the time node, ω is the scene variable reflecting randomness, and π ω Let Δt represent the probability of different scenarios occurring in stochastic programming, and Δt represent the unit time length.

[0017] Step 1.2, the energy storage aggregator sets the pricing strategy parameters (k) D-1 ,k D The upper limits of ) are respectively That is, it satisfies the following constraints:

[0018]

[0019]

[0020] Furthermore, step 2 specifically includes the following sub-steps:

[0021] Step 2.1: Based on historical wind power output data and meteorological forecast data within the system, the power market operator obtains the wind power output value and its probability for the future operating cycle. Using a two-stage stochastic programming approach, the wind power output is divided into three scenarios: high, medium, and low, with the probability distribution parameter set to π. ω ;

[0022] Step 2.2, the electricity market operator receives the pricing strategy parameters (k) from the energy storage aggregator. D-1 ,k D Based on the information, combined with the system topology and existing bidding data from other market participants, the lower-level modeling objective is to maximize total social welfare, i.e., minimize total power generation cost, as shown in the following expression:

[0023]

[0024] In the formula, i is the index of other thermal power generating units in the power system besides energy storage operators, and C iLet P be the variable cost per unit of power generation of the i-th generator unit. it Let be the output of the i-th generator unit at time t. and These represent the increased and decreased reserve capacity undertaken by the i-th generator unit at time t, respectively.

[0025] Step 2.3: The electricity market operator performs day-ahead-real-time market clearing modeling, considering day-ahead and real-time market power balance, upper and lower limits of thermal and wind turbine output, and energy storage capacity constraints, specifically including the following constraint expressions:

[0026]

[0027]

[0028]

[0029]

[0030]

[0031]

[0032]

[0033] In the formula, n is the index of a node in the system, j is the index of the load in the system, and L jt W is the magnitude of the j-th load at time t. t It is the output of the wind turbine at time t, Θ n This represents the set of nodes directly connected to node n, where m is the corresponding index, and B... nm W represents the branch admittance of the line between node n and node m. ωt θ represents the output of the wind turbine generator at time t in scenario ω. nt and θ nωt Let θ represent the voltage phase angle of node n at time t under market and operating conditions, respectively. mt and θ mωt Let represent the voltage phase angle of node m at time t under market and operating conditions, respectively. and P i For the upper and lower limits of the output of the i-th thermal power generator unit, SoC init For the initial capacity of energy storage, and SoC Let η represent the maximum and minimum energy storage capacities, and η be the energy storage charging and discharging efficiency parameter. The parameters in parentheses after the colon in the expression are the dual multipliers corresponding to the constraints, including: λ t , λ ωt , φt , and Δt is the unit time length.

[0034] Furthermore, step 3 specifically includes the following sub-steps:

[0035] Step 3.1: Rewrite the original problem of the lower-level model into a Lagrange dual form, and the constraints become:

[0036]

[0037]

[0038]

[0039]

[0040]

[0041]

[0042] Step 3.2: Construct the complementary relaxation conditions based on the KKT conditions, as shown in the following expression:

[0043]

[0044]

[0045]

[0046]

[0047]

[0048]

[0049]

[0050]

[0051]

[0052] Furthermore, in step 4, the binary expansion method is used to expand the variable representing the product of price and charge / discharge quantity in the objective function of the upper-level model. The processing procedure includes the following sub-steps:

[0053]

[0054]

[0055]

[0056]

[0057]

[0058] In the formula, k is an intermediate constant variable. For intermediate binary variables, for A constant within the range of values, For λ t The substitution variable is G, which is a constant and its value is greater than λ. t ten times.

[0059] Furthermore, in step 4, the process of linearizing the multiplication of the dual variable and the continuous variable in the complementary constraint using the Big M method includes the following sub-steps:

[0060]

[0061] In the formula, M is a constant. These are intermediate parameters for the Big M method.

[0062] Furthermore, step 5 specifically includes the following sub-steps:

[0063] The linearized single-layer model is input into the commercial optimization solver Gurobi. Initial and known parameter values ​​are set, and the optimal pricing strategy parameters are obtained through iterative calculation. and maximum market return W max and the resulting market price parameters and charge / discharge parameters Based on the results, the optimal bid for energy storage aggregators to participate in the day-ahead-real-time electricity market is determined.

[0064] Beneficial effects:

[0065] First, the method for energy storage aggregators to participate in day-ahead-real-time electricity market bidding in this invention focuses on the process of energy storage aggregators participating in day-ahead-real-time electricity market bidding. It uses the data transfer and iteration between aggregators and electricity market operators to model and express the mutual coupling and influence between aggregators and electricity market operators. It not only fully considers the maximum revenue of energy storage aggregators and the minimum cost requirements of electricity market operators, but also creatively transforms the two-layer model into a single-layer model through Lagrange duality and KKT conditions. This greatly simplifies the original complex iterative process between the upper and lower layers of the model, allowing it to be calculated using commercial optimization solvers, ensuring the solution speed while obtaining the globally optimal result.

[0066] Secondly, the method for energy storage aggregators to participate in day-ahead-real-time electricity market bidding in this invention takes into account the system fluctuations and uncertainties caused by the introduction of a large number of renewable energy sources into the current power system. Taking wind power as an example, it adopts a two-stage stochastic programming method to visualize the stochastic output process of wind power as random events with a fixed probability distribution. It considers all possible output scenarios in a way that minimizes the expected output cost. In the calculation, this not only avoids the modeling difficulties caused by parameter randomization but also yields the optimal expected cost result for the stochastic scenarios that power operators are concerned with.

[0067] Third, the method for energy storage aggregators to participate in day-ahead-real-time electricity market bidding in this invention addresses nonlinear terms arising in the modeling by employing different linearization schemes based on the different sources of nonlinearity. A binary expansion method is used to handle the product of price and energy storage charging / discharging capacity, and the Big M method is used to handle the product of dual and continuous variables in complementary relaxation conditions. This significantly reduces the difficulty of solving the model and facilitates the rapid generation of optimal bidding strategies. Attached Figure Description

[0068] Figure 1 This is a schematic diagram of the structure of a two-layer stochastic programming model for energy storage aggregators to participate in day-ahead-real-time market bidding, according to an embodiment of the present invention. Detailed Implementation

[0069] The following embodiments are provided to enable those skilled in the art to more fully understand the present invention, but do not limit the invention in any way.

[0070] Bilevel optimization, through parameter passing and iteration, mathematically rigorously captures the interaction between the strategic decisions of self-interested participants in the upper-level model and the competitive market clearing in the lower-level model. It captures the strategic, profit-driven behavior of self-interested market participants and identifies the market outcomes arising from their interactions, making it a crucial tool for solving interactive competitive strategy problems. Stochastic programming, a branch of planning problems, addresses decision-making problems under uncertainty by defining the system's uncertainty into several typical scenarios with known probability distributions, and then achieving the expected optimal solution through a multi-stage optimization process. Combining bilevel optimization and stochastic programming effectively addresses market bidding scenarios involving multiple actors and at multiple market scales. It not only achieves expected optimality under stochastic conditions but also maximizes the returns of market participants, fully stimulating market vitality.

[0071] See Figure 1This invention discloses a method for energy storage aggregators to participate in day-ahead-real-time electricity market bidding. The upper-level model in the two-layer model aims to maximize the market revenue of the energy storage aggregator, while the lower-level model aims to achieve day-ahead-real-time electricity market clearing through stochastic programming while maximizing total social welfare. The upper and lower-level models involve real-time parameter transfer and iteration. The specific steps are as follows:

[0072] Step 1: Energy storage aggregators obtain cost information and derive the unit power (kWh) charge / discharge cost function C. bat It simultaneously provides electrical energy and reserves to participate in the day-ahead and real-time electricity markets, respectively, with proposed bidding strategy parameters k. D-1 and k D And pass it to the lower layer, given the range of pricing strategy parameters, and obtain the lower layer price parameter (λ). t ,λ ωt and charge / discharge parameters Based on this, the upper-level model is built with the goal of maximizing the market participation return W.

[0073] Step 2: The goal of electricity market operators is to simultaneously clear the energy and reserve markets while minimizing generation and consumption costs. Obtain the pricing strategy parameters (k) of energy storage aggregators. D-1 ,k D Following this, the lower-level model formulates the total generation cost equation, completes the system power flow constraints using stochastic programming, takes into account the generation and reserve output constraints of energy storage aggregators and other generators within the system, and completes the lower-level model modeling. The output results include the day-ahead market clearing price λ. t and real-time market clearing price λ ωt Planned charging volume of energy storage aggregators and discharge quantity Energy storage aggregators reduce reserve requirements and reduce reserve The result is then returned to the upper layer.

[0074] Step 3: Rewrite the objective function and constraints in the original problem (PL) of the lower-level model into the Lagrange dual form (DL), and use KKT conditions to replace the lower-level model with the original problem feasibility conditions, the dual problem feasibility conditions, and complementary relaxation constraints.

[0075] Step 4: Based on Step 3, the two-layer model is further transformed into a single-layer model. In the transformed single-layer model, the objective function is the same as that of the original upper-layer model, and the constraints include the constraints of the original upper-layer model, the feasibility conditions of the original problem in the lower-layer model, the feasibility conditions of the dual problem in the lower-layer model, and the complementary relaxation constraints of the lower-layer problem.

[0076] Step 5: Linearize the product of dual and continuous variables in the complementary constraints using the Big M method; and linearize the product of price and charge / discharge quantities in the objective function of the upper-level model using the binary expansion method. and Process it.

[0077] Step 6: Based on the above steps, solve the linearized single-layer model using the commercial solver Gurobi to obtain the optimal pricing strategy parameters. and maximum market return W max and the resulting market price parameters and charge / discharge parameters

[0078] In this embodiment, the power system topology adopts the IEEE RTS24 standard node, which includes a total of 24 nodes, 8 thermal power units, 1 wind power unit, 1 energy storage aggregator, and 15 load nodes. The wind power unit is located at node 1, and the energy storage aggregator is located at node 15. The energy storage aggregator's pricing strategy has an upper limit. The probability distribution for high, medium, and low wind power output scenarios is (0.5, 0.3, 0.2), where (10, 10).

[0079] This embodiment is carried out according to the following steps:

[0080] Step 1: Complete the upper-level modeling with the goal of maximizing the market participation return W. This specifically includes:

[0081] Step 1.1, the energy storage aggregator obtains its unit power (kWh) charge / discharge cost function C. bat According to its pricing strategy parameter (k) D-1 ,k D ), respectively according to k D-1 ·C bat and k D ·C bat Price quotes are provided for two usage scenarios: electrical energy and backup power, given the known lower-level price parameter (λ). t ,λ ωt and charge / discharge parameters Based on this, its market participation returns are expressed as follows:

[0082]

[0083] In the formula, t represents the time node, ω is the scene variable reflecting randomness, and π ω Let Δt represent the probability of different scenarios occurring in stochastic programming, and let Δt represent the unit time length.

[0084] Step 1.2, the energy storage aggregator sets the pricing strategy parameters (k) D-1 ,kD The upper limits of ) are respectively That is, it satisfies the following constraints:

[0085]

[0086]

[0087] Step 2, complete the lower-level model modeling. This specifically includes:

[0088] Step 2.1: Based on historical wind power output data and meteorological forecast data within the system, the power market operator obtains the wind power output value and its probability for the future operating cycle. Using a two-stage stochastic programming approach, the wind power output is divided into three scenarios: high, medium, and low, with the probability distribution parameter set to π. ω .

[0089] Step 2.2, the electricity market operator receives the pricing strategy parameters (k) from the energy storage aggregator. D-1 ,k D Based on the information, combined with the system topology and existing bidding data from other market participants, the lower-level modeling objective is to maximize total social welfare, i.e., minimize total power generation cost, as shown in the following expression:

[0090]

[0091] In the formula, i is the index of other thermal power generating units in the power system besides energy storage operators, and C i Let P be the variable cost per unit of power generation of the i-th generator unit. it Let be the output of the i-th generator unit at time t. and These represent the increased and decreased reserve capacity undertaken by the i-th generator unit at time t, respectively.

[0092] Step 2.3: The electricity market operator performs day-ahead-real-time market clearing modeling, considering day-ahead and real-time market power balance, upper and lower limits of thermal and wind turbine output constraints, energy storage capacity constraints, etc., specifically including the following constraint expressions:

[0093]

[0094]

[0095]

[0096]

[0097]

[0098]

[0099]

[0100] In the formula, n is the index of a node in the system, j is the index of the load in the system, and L jt W is the magnitude of the j-th load at time t. t It is the output of the wind turbine at time t, Θ n This represents the set of nodes directly connected to node n, where m is the corresponding index, and B... nm W represents the branch admittance of the line between nodes n and m. ωt θ represents the output of the wind turbine generator at time t in scenario ω. nt and θ nωt Let represent the voltage phase angle of node n at time t under market and operating conditions, respectively. and P i For the upper and lower limits of the output of the i-th thermal power generator unit, SoC init For the initial capacity of energy storage, and SoC Let represent the maximum and minimum energy storage capacities, and η represent the energy storage charging and discharging efficiency parameter. The parameters in parentheses after the colon in the expression are the dual multipliers corresponding to the constraints.

[0101] Step 3: Rewrite the objective function and constraints in the original problem (PL) of the lower-level model into Lagrange dual form (DL). Specifically, this includes:

[0102] Step 3.1: Based on the lower-level modeling completed in Step 2, rewrite the original problem (PL) in its Lagrange dual form, and the constraints become:

[0103]

[0104]

[0105]

[0106]

[0107]

[0108]

[0109] Step 3.2, further, based on the KKT conditions in step 2, construct the complementary relaxation conditions, as shown in the following expression:

[0110]

[0111]

[0112]

[0113]

[0114]

[0115]

[0116]

[0117]

[0118]

[0119] Step 4: Based on the transformation of the lower-level model's primal problem (PL) into its dual problem (DL) in Step 3, the lower-level model is replaced with the primal problem's feasibility conditions, the dual problem's feasibility conditions, and complementary relaxation conditions according to the KKT conditions. This transforms the two-layer model into a single-layer model. The specific structure of the transformed single-layer model is as follows:

[0120] The objective function is the objective function of the upper-level model;

[0121] The constraints include upper-level model constraints, lower-level model primal problem feasibility conditions, lower-level model dual problem feasibility conditions, and complementary relaxation conditions.

[0122] Step 5: The Big M method is used to linearize the multiplication of the dual variable and the continuous variable in the complementary constraint conditions. The binary expansion method is used to linearize the multiplication of price and charge / discharge quantity in the objective function of the upper-level model. and The process will be handled. Specifically, this includes:

[0123] Step 5.1, for the price variable (λ) in the objective function of the upper-level model t ,λ ωt and energy storage charging and discharging capacity The product is linearized using a binary expansion method, so that... The following example illustrates the detailed steps; other similar product terms are handled in the same way.

[0124]

[0125]

[0126]

[0127]

[0128]

[0129] In the formula, k is an intermediate constant variable. For intermediate binary variables, for A constant within the range of values, For λ t The substitution variable, G, is a sufficiently large constant, typically taken to be greater than λ. t Ten times is ideal.

[0130] Step 5.2: Linearize the nonlinearity caused by the product of dual and continuous variables in the complementary constraints using the Big M method:

[0131]

[0132] In the formula, M is a sufficiently large constant. These are intermediate parameters for the Big M method.

[0133] Step 6: Based on steps 1-5, input the linearized single-layer model into the commercial optimization solver Gurobi, set the initial parameters and known parameter values, and obtain the optimal pricing strategy parameters through iterative calculation. and maximum market return W max and the resulting market price parameters and charge / discharge parameters Based on the results, the optimal bid for energy storage aggregators to participate in the day-ahead-real-time electricity market is determined.

[0134] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A method for energy storage aggregators to participate in day-ahead-real-time electricity market bidding, characterized in that, The electricity market bidding method includes the following steps: Step 1: The energy storage aggregator obtains cost information and derives the unit power charge / discharge cost function. Energy storage aggregators will provide electricity and backup services in the day-ahead and real-time electricity markets, respectively, and will formulate pricing strategy parameters. and The data is passed down to the lower level, given the range of pricing strategy parameters, to obtain the lower-level price parameters. and charge / discharge parameters Based on the benefits of participating in the market Maximize the completion of upper-level modeling for the objective; Step 2: Obtain the pricing strategy parameters of the energy storage aggregator. ,in and Based on the day-ahead and real-time market pricing strategy parameters, a total generation cost equation is formulated. System power flow constraints are handled using stochastic programming, taking into account the generation and reserve output constraints of energy storage aggregators and thermal and wind turbine units within the system. The lower-level model is then completed, and its output includes the day-ahead market clearing price. and real-time market clearing price Planned charging volume of energy storage aggregators and discharge quantity Energy storage aggregators reduce reserve capacity and increase reserve The results are then returned to the upper-level model. Step 3: Rewrite the objective function and constraints in the original problem of the lower-level model into Lagrange dual form, and use KKT conditions to replace the lower-level model with the original problem feasibility conditions, dual problem feasibility conditions, and complementary relaxation constraints, thus transforming the two-level model into a single-level model. In the transformed single-level model, the objective function is the same as the original upper-level model objective function, and the constraints include the original upper-level model constraints, the original problem feasibility conditions of the lower-level model, the dual problem feasibility conditions of the lower-level model, and the complementary relaxation constraints of the lower-level model. Step 4: The Big M method is used to linearize the product term of the dual and continuous variables in the complementary relaxation constraints. The binary expansion method is used to linearize the product term of price and charge / discharge in the objective function of the upper-level model. , , and Process it; Step 5: Solve the linearized single-layer model using the commercial solver Gurobi to obtain the optimal pricing strategy parameters. and maximum market return and the resulting market price parameters and charge / discharge parameters .

2. The method for energy storage aggregators to participate in day-ahead-real-time electricity market bidding according to claim 1, characterized in that, Step 1 specifically includes the following sub-steps: Step 1.1, the energy storage aggregator obtains its unit power charge / discharge cost function. According to its pricing strategy parameters According to and Quotations are provided for both electrical energy and backup usage scenarios, with known lower-level price parameters. and charge / discharge parameters Based on this, its market participation benefits The expression is as follows: ； In the formula, Represents a time point. To reflect the randomness of scene variables, Let represent the probability of different scenarios occurring in stochastic programming. The unit of time length; Step 1.2, Energy storage aggregators set pricing strategy parameters The upper limits are respectively That is, it satisfies the following constraint relationship: ； 。 3. The method for energy storage aggregators to participate in day-ahead-real-time electricity market bidding according to claim 1, characterized in that, Step 2 specifically includes the following sub-steps: Step 2.1: Based on historical wind power output data and meteorological forecast data within the system, the power market operator obtains the wind power output value and its probability for the future operating cycle. Using a two-stage stochastic programming approach, the wind power output is divided into three scenarios: high, medium, and low. The probability distribution parameters are set as follows: ; Step 2.2: The electricity market operator receives the pricing strategy parameters from the energy storage aggregator. Based on the information, combined with the system topology and existing bidding data from other market participants, the lower-level modeling objective is to maximize total social welfare, i.e., minimize total power generation cost, as shown in the following expression: ； In the formula, This serves as an index for thermal power generating units in the power system, excluding energy storage operators. For the first Variable cost per unit of electricity generated by a thermal power generating unit For the first Each thermal power generating unit at time of effort, and The first Each thermal power generating unit at time The responsibility for increasing and decreasing reserve capacity; Step 2.3: The electricity market operator performs day-ahead-real-time market clearing modeling, considering day-ahead and real-time market power balance, upper and lower limits of thermal and wind turbine output constraints, and energy storage capacity constraints, specifically including the following constraint expressions: ； ； ； ； ； ； ； In the formula, It is the index of a node in the system. It is an index of the load in the system. It is the first The load at time Size, It is the wind turbine generator set at all times of effort, Represents nodes The set of directly connected nodes It corresponds to the index. Represents a node and nodes Branch admittance of the line between them Indicates wind turbine generator set in the scene Next moment of effort, Indicates the wind turbine generator at time The upper limit of output, and Representing nodes under market and operating conditions respectively. At any moment voltage phase angle, and Representing nodes under market and operating conditions respectively. At any moment voltage phase angle, and For the first The upper and lower limits of the output of each thermal power generating unit For the initial capacity of energy storage, and For the maximum and minimum capacity of energy storage, The efficiency parameters for energy storage charging and discharging are the parameters in parentheses after the colon in the expression, including... , , , , , , , , , , The dual multipliers corresponding to the constraints. The unit of time is the length of time.

4. The method for energy storage aggregators to participate in day-ahead-real-time electricity market bidding according to claim 1, characterized in that, Step 3 specifically includes the following sub-steps: Step 3.1: Rewrite the original problem of the lower-level model into a Lagrange dual form, and the constraints become: ； ； ； ； ； ； In the formula, For the first Variable cost per unit of electricity generated by a thermal power generating unit The unit of time length , , , , , , , , , and The dual multipliers corresponding to the constraints. To reflect the randomness of scene variables, Let represent the probability of different scenarios occurring in stochastic programming. These are the efficiency parameters for energy storage charging and discharging. This is the total calculation time period; Step 3.2: Construct the complementary relaxation constraints based on the KKT conditions, as shown in the following expression. As shown in the formula: ； ； ； ； ； ； ； ； ； In the formula, and For the first The upper and lower limits of the output of each thermal power generating unit For the first Each thermal power generating unit at time of effort, and The first Each thermal power generating unit at time The responsibility for increasing and decreasing reserve capacity. It is the wind turbine generator set at all times of effort, Indicates the wind turbine generator at time The upper limit of output, For the initial capacity of energy storage, and These represent the maximum and minimum energy storage capacities.

5. The method for energy storage aggregators to participate in day-ahead-real-time electricity market bidding according to claim 1, characterized in that, In step 4, the binary expansion method is used to expand the variable representing the product of price and charge / discharge quantity in the objective function of the upper-level model. The processing procedure includes the following sub-steps: ； ； ； ； ； In the formula, For intermediate constant variables, For intermediate binary variables, for A constant within the range of values, for Alternative variables, It is a constant, and its value is greater than ten times.

6. The method for energy storage aggregators to participate in day-ahead-real-time electricity market bidding according to claim 1, characterized in that, Step 4, the process of linearizing the product term of the dual and continuous variables in the complementary relaxation constraints using the Big M method, includes the following sub-steps: ； In the formula, It is a constant. These are intermediate parameters for the Big M method. The dual multipliers corresponding to the constraints. and For the first in the system The upper and lower limits of the output of each thermal power generating unit.

7. The method for energy storage aggregators to participate in day-ahead-real-time electricity market bidding according to claim 1, characterized in that, Step 5 specifically includes the following sub-steps: The linearized single-layer model is input into the commercial optimization solver Gurobi. Initial and known parameter values ​​are set, and the optimal pricing strategy parameters are obtained through iterative calculation. and maximum market return and the resulting market price parameters and charge / discharge parameters Based on the results, the optimal bid for energy storage aggregators to participate in the day-ahead-real-time electricity market is determined.

Images