A multi-time scale market bidding method and system for energy storage systems considering dynamic opportunity cost

Abstract

This invention discloses a multi-timescale market bidding method and system for energy storage systems considering dynamic opportunity costs, relating to the fields of electricity market and energy storage optimization technology. Through a dynamic opportunity cost quantification model, the potential impact of current decisions on future returns is dynamically quantified as constraints. While avoiding short-sighted decisions, a multi-market time-series coupling mechanism is constructed using the capacity state transfer equation and cross-market power balance constraints. This ensures the physical feasibility and temporal consistency of the dynamic allocation of energy storage capacity across day-ahead, intraday, and real-time markets. Furthermore, a stochastic-robust hybrid optimization framework is proposed, integrating generative adversarial network scenario generation, conditional risk value, and robustness margin to unify the optimization of expected returns and extreme risks. The constraints are transformed into optimization problems using KKT conditions and the Big M method, and solved using the ADMM distributed algorithm, ensuring the model's feasibility and real-time performance in engineering practice.

Description

Technical Field

[0001] This invention relates to the field of electricity market and energy storage optimization technology, and in particular to a multi-timescale market bidding method and system for energy storage systems that considers dynamic opportunity costs. Background Technology

[0002] The integration of a high proportion of renewable energy into the grid has become an inevitable trend in power grid development. The randomness and volatility of wind and solar power output pose a severe challenge to the real-time balancing capabilities of the power system. Against this backdrop, energy storage systems, with their rapid and flexible adjustment characteristics, are considered a key technological means to ensure grid security and promote the consumption of new energy sources. Promoting the participation of energy storage systems in electricity trading through market mechanisms is not only an important way to realize their own commercial value, but also a core element in improving the overall operating efficiency and economy of the power system.

[0003] Currently, the electricity market has gradually formed a multi-timescale trading system covering day-ahead, intraday, and real-time. However, energy storage systems face fundamental technical challenges in making optimal decisions in this complex market environment: on the one hand, energy storage charging and discharging activities consume limited capacity resources, and current decisions directly affect the ability to participate in higher-value market opportunities in the future. This "dynamic opportunity cost" is difficult to characterize using traditional static cost models. On the other hand, there is a close temporal coupling relationship between markets at different timescales. How to effectively coordinate the long-term nature of day-ahead planning, the flexibility of intraday adjustments, and the speed of real-time response in decision-making, and avoid decision conflicts and capacity waste, is a problem that current technical solutions have not yet solved well. Therefore, developing a bidding strategy that can accurately quantify dynamic opportunity costs and achieve multi-market collaborative optimization has become an urgent need for the industry.

[0004] Traditional methods generally employ static cost models and centralized optimization architectures, which are ill-suited to the dynamic and uncertain nature of the electricity market. Existing technology 1 (CN202410076643.8) discloses a method for energy storage aggregators to participate in day-ahead and real-time electricity market bidding, representing a current mainstream technical solution. This method achieves coordinated optimization of the day-ahead and real-time markets by constructing a two-layer optimization framework. Specifically, this scheme first requires aggregating distributed energy storage resources and collecting global data, including total capacity and historical bidding information. Then, at the upper layer, a stochastic programming method is used to formulate the day-ahead market's charging and discharging plan, with its objective function optimizing expected revenue in a linear programming form, while considering power balance constraints and energy storage operation limitations. At the lower layer, a model predictive control (MPC) algorithm is used to continuously adjust the real-time market strategy to cope with fluctuations in renewable energy output and load forecasting deviations. The device implementation of this method relies on a unified scheduling architecture of a central controller, requiring a high-performance centralized data processor to achieve global information integration, and a standardized market interface module to complete bidding applications for markets at multiple time scales. However, its technical solution uses fixed values ​​to characterize energy storage loss costs, and fails to establish a correlation model between opportunity costs and dynamically changing market prices and wind and solar power output. This results in the system being unable to assess the potential impact of current decisions on future revenues, and it also ignores intraday market connections, resulting in insufficient capacity utilization and a lack of robust mechanisms to cope with extreme fluctuations.

[0005] Existing technology 2 (CN202411265115.3) discloses a decentralized multi-element energy storage model aggregation method considering dynamic boundary adjustment, which proposes an improved scheme from the perspective of energy storage resource aggregation. Its technical solution first performs dynamic clustering analysis based on the SOC state of each energy storage unit to generate energy storage cluster aggregation boundaries with different adjustment capabilities. Then, it uses a quadratic programming method to dynamically adjust the charge and discharge power thresholds of each cluster. In the day-ahead market phase, adjustable capacity is uniformly declared according to the aggregation boundaries, while in the real-time market phase, the scheduling instructions are decomposed to the unit level for execution through an instruction distributor. This scheme innovatively designs a boundary calculation engine in the device structure, which can update the dynamic characteristic parameters of the energy storage cluster in real time. It also uses a linear decay model (multiplying the number of cycles by a fixed decay coefficient) to simplify the estimation of lifetime loss. Although this reduces computational complexity, it fails to fully consider the nonlinear impact of different charge and discharge depths on battery life, resulting in bias in opportunity cost estimation. Furthermore, its centralized solution mode of the optimization architecture has a slow response speed, making it difficult to meet the high-frequency demand of minute-level decision-making in the real-time market. It also has shortcomings in the architecture design for multi-market collaborative optimization, resulting in high cross-market revenue losses.

[0006] In summary, while these existing technologies can enable energy storage to participate in the market in specific scenarios, they still face key challenges when dealing with the strong uncertainties brought about by the high proportion of renewable energy grid connection. These challenges include inaccurate quantification of opportunity costs, insufficient coordination across multiple time scales, and inadequate computational efficiency. Systematic innovation in model building, optimization methods, and device implementation is urgently needed. To address this, a multi-time-scale market bidding method and system for energy storage systems that considers dynamic opportunity costs is proposed. Summary of the Invention

[0007] The main objective of this invention is to provide a multi-timescale market bidding method and system for energy storage systems that considers dynamic opportunity costs. By constructing a dynamic opportunity cost quantification model and a multi-layer nested optimization framework, it achieves synergistic optimization and profit maximization of energy storage in day-ahead, intraday, and real-time electricity markets. At the same time, by combining a stochastic-robust hybrid optimization method, it effectively balances expected market returns and extreme volatility risks, significantly improving the decision-making quality and economic efficiency of energy storage participating in multi-timescale markets, and effectively solving the problems in the background technology.

[0008] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0009] A multi-timescale market bidding method for energy storage systems that considers dynamic opportunity costs includes the following steps:

[0010] S1: Acquire multi-source data from the electricity market and energy storage systems, including real-time electricity prices, grid load, energy storage state of charge (SOC), energy storage power, and renewable energy output data, and combine them with a pre-set prediction model to form a complete dataset containing the current state and future predictions.

[0011] S2: Generative Adversarial Network (GAN) is used to simulate and generate various new energy output scenarios and their probabilities, and dynamic opportunity costs are evaluated based on discount theory.

[0012] Specifically:

[0013] To describe the randomness of renewable energy output, a generative adversarial network (GAN) is used to generate a large number of possible high, medium, and low output scenarios and their probability distributions. The minimax game objective function of the GAN is defined as follows:

[0014]

[0015] in, As a generator, it will generate a random noise vector Mapped to output scenarios; It serves as a discriminator, used to determine the authenticity of the input scene; Data distribution contributing to the true history;

[0016] After the Generative Adversarial Network (GAN) is trained, representative output scenarios are obtained by clustering and reducing the generated output scenarios. and the scene of output Corresponding probability of occurrence ,in .

[0017] Dynamic opportunity cost is defined as the sum of the present values ​​of potential future gains lost due to current decisions, and is calculated using the following explicit quantification model:

[0018]

[0019] in, For time period Dynamic opportunity cost; It is a discount factor that reflects the time value of money; For time period The expected maximum return, through the scenario and its probability Perform calculations; For time period The actual achievable benefits take into account the impact of the current decision on energy storage capacity utilization; this explicit quantification model is used to dynamically quantify the impact of the current decision on future potential benefits into cost constraint terms that can be embedded in the collaborative optimization model, by calculating the obtained... Compared with dynamically updated data based on historical market data and capacity constraints Compare and generate constraints And embed it into the subsequent optimization model to avoid short-sighted decision-making.

[0020] S3: By using the capacity state transfer equation and cross-market power balance constraints, the SOC and charging / discharging power of the energy storage system are dynamically correlated among the day-ahead market, intraday market, and real-time market. A multi-timescale market collaborative optimization model integrating dynamic opportunity cost constraints is constructed, and the evaluated dynamic opportunity cost is embedded as a constraint condition into the collaborative optimization model.

[0021] Specifically:

[0022] To achieve multi-timescale coordination across different markets, the following constraints are used to dynamically correlate the state and power of the energy storage system across the day-ahead market, intraday market, and real-time market, ensuring the temporal consistency of decision-making.

[0023] After the day-ahead market closes and before the intraday market begins, the initial electricity available for the intraday market is equal to the remaining electricity after the day-ahead market plan has been executed. Therefore, the state transfer equation from the day-ahead market to the intraday market is as follows:

[0024]

[0025] in For the current moment The initial electricity volume flowing from the previous day's market into the intraday market serves as the starting point for intraday market decisions. For the current moment Current market status of charge of energy storage; The discharge efficiency of the energy storage system; For the current moment The market's net power at present; This refers to the trading interval in the market prior to the current day;

[0026] Similarly, before the real-time market begins, its available electricity depends on the execution results of the intraday market, hence the state transfer equation from the intraday market to the real-time market:

[0027]

[0028] in For the current moment Initial electricity volume status flowing from the intraday market to the real-time market; The charging efficiency of the energy storage system; For the current moment Charging power in the market during the day;

[0029] In each time period of the real-time market, the SOC changes according to the real-time charging and discharging power, hence the state evolution equation within the real-time market is as follows:

[0030]

[0031] in for Real-time market status of energy storage charge; For the current moment Real-time market discharge power; To implement the market's trading time interval;

[0032] To ensure the safe operation of energy storage devices, the total power declared in three markets at the same time cannot exceed their physical limits. The power symbol convention is that discharge is positive and charging is negative, hence there are cross-market power balance constraints.

[0033]

[0034] in, , , These represent the energy storage system's performance in the day-ahead market, intraday market, and real-time market at the current moment. The declared net power.

[0035] In the above equation , , The decision variables, their range of values, and their interrelationships are constrained by dynamic opportunity cost. The direct impact is that a high opportunity cost means that occupying capacity in the current period is costly, and the optimization model will tend to reduce power declarations in that market, thus affecting state transmission.

[0036] To establish an optimization framework capable of coordinating markets across multiple time scales while considering dynamic opportunity costs and market equilibrium, the objective function of the co-optimization model is defined as a stochastic-robust hybrid form, employing a three-level nested structure. The objective function of the co-optimization model is defined as follows:

[0037]

[0038] in, For expected returns, In the scene The following benefits; Conditional risk value, , All are auxiliary variables; For robust margin, It is a set of uncertainties constructed based on Wasserstein distance, used to guard against errors in the probability distribution itself; Risk aversion coefficient; This is the robustness weighting coefficient.

[0039] The three-level nested structure specifically includes:

[0040] The upper-level problem aims to maximize the overall benefits of energy storage systems across multiple time scales by optimizing the pricing strategy parameters for each market.

[0041] For mid-level problems, the dynamic opportunity cost is calculated in real time, and constraints are generated. Feedback to higher levels regarding the issue, including The preset opportunity cost threshold;

[0042] The lower-level problem simulates the market clearing process with the goal of minimizing total social cost. The total social cost includes the power generation cost of conventional generator sets and the comprehensive cost of energy storage systems. The comprehensive cost of energy storage systems includes operating costs and dynamic opportunity costs.

[0043] The objective function, which aims to maximize the overall returns of energy storage systems across multiple time scales, is defined as follows:

[0044]

[0045] The constraints of the objective function, which aims to maximize the overall returns of energy storage systems across multiple time scales, are defined as follows: ;

[0046] in, , , These are the pricing strategy parameters for energy storage systems in the day-ahead market, intraday market, and real-time market, respectively. , , These represent the maximum values ​​of the pricing strategy parameters for energy storage systems in the day-ahead market, intraday market, and real-time market, respectively. , , These refer to the daytime market, intraday market, and real-time market at the current moment. Electricity price; For the current moment The overall cost of energy storage systems;

[0047] The objective function that aims to minimize total social cost is defined as follows:

[0048]

[0049] The constraints of the objective function aimed at minimizing total social cost include at least power balance constraints, generator output constraints, and energy storage operation constraints.

[0050] The power balance constraint is defined as follows: ;

[0051] The generator set output constraint is defined as follows: ;

[0052] The energy storage operation constraints are defined as follows:

[0053] ;

[0054] in, For the first The power generation cost function of a generator set is used to describe its output. The relationship with cost can be expressed as: , , , All are unit characteristic coefficients; This is the overall cost function of the energy storage system; For the output of energy storage systems; For the first The output of the generator set; For the energy storage system at the current moment Net charge and discharge power, This indicates that the energy storage system is in a discharging state. This indicates that the energy storage system is in a charging state. This indicates that the energy storage system is in an idle state; Contribute to the total output of renewable resources; For nodes Load demand; , These represent the number of generator sets and load nodes, respectively. , The first The lower and upper limits of the output of the generator set; , These are the lower and upper limits of the SOC (State of Charge) for energy storage systems, respectively. , These represent the lower and upper limits of the energy storage system's output, respectively.

[0055] S4: Using the Karush-Kuhn-Tucker condition and the alternating direction multiplier method, the collaborative optimization model is transformed into a distributed optimization problem, and the collaborative bidding strategies for the day-ahead market, intraday market and real-time market are obtained by solving the problem.

[0056] Specifically:

[0057] S41: The lower-level problem is transformed into a dual form using the Karush-Kuhn-Tucker condition, and the complementary relaxation condition is linearized using the Big M method, transforming the three-layer nested collaborative optimization model into a single-layer mixed integer linear programming (MILP) problem.

[0058] The conversion steps specifically include:

[0059] S411: Construct the Lagrange function and introduce Lagrange multipliers (dual variables). , ;

[0060] S412: To ensure the original feasibility, all constraints of the lower-level problem must be satisfied:

[0061] ;

[0062] S412: To ensure duality feasibility, the following must be satisfied:

[0063] With inequality constraints The relevant non-negative Lagrange multipliers: , ;

[0064] The gradient condition is:

[0065] ;

[0066] The complementary relaxation condition is: .

[0067] S413: For nonlinear complementary relaxation conditions, introduce binary variables. and large constants Linearize it:

[0068] .

[0069] By following the above transformation steps, the three-level nested collaborative optimization model can be transformed into a single-level mixed integer linear programming (MILP) problem.

[0070] S42: The transformed single-level mixed integer linear programming (MILP) problem is solved in a distributed manner using the alternating direction multiplier method. Specifically, the global optimization variable is decomposed into subproblems corresponding to the day-ahead market, intraday market, and real-time market. , and Parallel solutions are obtained and coordinated through global consistency constraints, iteratively updating dual variables until convergence.

[0071] The solution steps are as follows:

[0072] S421: Solving for local variables:

[0073] ;

[0074] S422: Matrix form of global variable consistency constraints:

[0075] ;

[0076] S423: Dual variable update: ;

[0077] S424: Algorithm convergence condition: Iterate to the original residual and dual residuals When the time is sufficiently short, the algorithm converges.

[0078] S5: Perform security verification on the generated bidding strategy to ensure that it complies with the operating constraints of the energy storage system;

[0079] Specifically:

[0080] Safety verification includes at least hard constraint verification of the energy storage system's state of charge (SOC), charging and discharging power, and power change rate, expressed as:

[0081]

[0082] in, This represents the upper limit of the energy storage ramp-up rate.

[0083] S6: Executes strategies in a rolling time-domain manner and makes incremental optimizations and adjustments based on the latest market data and system status.

[0084] Specifically:

[0085] The model runs in a rolling time-domain manner. At each decision point... It only executes the strategy for the current time period and updates the prediction information set based on the latest observation data, then re-solves the optimization problem. When drastic fluctuations are detected, incremental optimization is triggered, re-solving only the affected time period, significantly improving computational efficiency.

[0086] A multi-timescale market bidding system for energy storage systems that considers dynamic opportunity costs includes:

[0087] The data acquisition layer is used to acquire or predict market electricity prices, grid load, renewable energy output, and energy storage system status data in real time.

[0088] An optimized computing layer, connected to the data acquisition layer, includes a scene generation and classification unit and a core optimization solution engine;

[0089] The scene generation and classification unit is used to generate uncertain scenes based on GAN;

[0090] The core optimization engine embeds the dynamic opportunity cost quantification model and is used to execute the steps of the method described above, construct and solve the collaborative optimization model, and generate a bidding strategy.

[0091] The strategy execution layer, connected to the optimization calculation layer, includes a dynamic opportunity cost feedback module, a security verification module, and an energy storage device control terminal;

[0092] The dynamic opportunity cost feedback module is used to feed the calculated dynamic opportunity cost value back to the optimization model;

[0093] The security verification module is used to perform compliance checks on the generated bidding strategies;

[0094] The control terminal of the energy storage device is used to convert the bidding strategy that has been verified for security into control commands and send them out for execution.

[0095] The present invention has the following beneficial effects:

[0096] Compared with existing technologies, this solution maximizes global revenue across day-ahead, intraday, and real-time markets through a dynamic opportunity cost model and a three-layer nested optimization framework. Simulation results show that the overall revenue of the energy storage system is increased by 28%, and the cross-market capacity utilization rate is increased by 35%.

[0097] Compared with existing technologies, this scheme reduces the volatility of returns by 40% and the tail risk (CVaR) by 55% under the multi-stage stochastic-robust hybrid optimization framework in extreme scenarios (such as a sudden drop of 50% in wind power output).

[0098] Compared with existing technologies, this solution effectively suppresses short-sighted decision-making through explicit opportunity cost constraints, reducing future revenue losses caused by energy storage capacity occupancy by 32%.

[0099] Compared with existing technologies, this solution reduces the solution time by 45% and achieves a real-time market strategy adjustment response speed in minutes by using the alternating direction multiplier method to transform the complex three-layer model into a single-layer MILP problem.

[0100] Compared with existing technologies, this solution can balance the safety and economy of energy storage systems, making the violation rate of the system's dynamic energy storage capacity constraint (20%≤SOC≤90%) close to 0%; the unit output cost is reduced by 24%, and the charge and discharge cycle loss cost is reduced by 18%. Attached Figure Description

[0101] Figure 1 This is a schematic diagram of the multi-timescale market bidding system architecture for energy storage systems proposed in this invention;

[0102] Figure 2 This is a schematic diagram illustrating the decision-making process of the energy storage system proposed in this invention participating in market bidding across multiple time scales.

[0103] Figure 3 This is a convergence verification diagram of the alternating direction multiplier algorithm proposed in this invention. Detailed Implementation

[0104] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0105] See Figure 1 The diagram shown is an architecture diagram of a multi-timescale market bidding system for energy storage provided by the present invention.

[0106] The system adopts a layered design, divided from top to bottom into a data acquisition layer, an optimization calculation layer, and a strategy execution layer. Specifically:

[0107] The data acquisition layer serves as the foundation, acquiring real-time data through various interface modules: the sensor interface directly connects to grid sensors to collect real-time electricity price and load data; the load forecasting module forecasts short-term load based on historical data and weather information; and the energy storage power detection module monitors the operating status of the energy storage system in real time, including current SOC and charging / discharging power.

[0108] Among them, regarding the future Predictions for specific time periods are essentially based on historical information sets. Conditional expectation estimation. Taking electricity price forecasting as an example, in one possible implementation, its model can be expressed as:

[0109]

[0110] in Indicated as in Always The predicted electricity price at any given time. , , , These represent the day-to-day market, intraday market, and real-time market, respectively. Pre-trained prediction functions, such as LSTM neural networks; This is the prediction error term.

[0111] It should be noted that the prediction models for load and renewable energy output are similar in form, and the output results are complete prediction trajectories, which together constitute the parameter set of the optimization problem.

[0112] The optimization computing layer serves as the core decision-making hub of the system. The data storage and analysis unit cleans, archives, and extracts features from the massive amounts of data uploaded by the acquisition layer. The scene classification unit uses techniques such as generative adversarial networks (GANs) to generate uncertain scenes based on wind and solar power output predictions, providing input for stochastic-robust optimization.

[0113] The strategy execution layer is responsible for the final issuance of instructions and safety control. The dynamic opportunity cost feedback module feeds back the real-time opportunity cost value obtained by the calculation layer to the optimization model. The safety verification module performs compliance checks on the generated charging and discharging strategies to ensure that the SOC is always within the safe range and the power change rate is within the equipment limit. Finally, the safe and error-free instructions are converted into control signals through the energy storage device control terminal to drive the energy storage converter to perform precise charging and discharging actions.

[0114] This system architecture realizes a closed-loop automated process from data perception to intelligent decision-making and then to secure execution.

[0115] See Figure 2 The charging decision flowchart provided by the present invention illustrates the decision-making process of an energy storage system participating in multi-timescale market bidding, including: real-time status acquisition (electricity price, SOC, load, etc.); future load and electricity price prediction; multi-stage stochastic-robust hybrid optimization; dynamic opportunity cost feedback machine; alternating direction multiplier (ADMM) distributed algorithm solution; and collaborative bidding in day-ahead, intraday, and real-time markets.

[0116] Figure 3The convergence verification diagram of the ADMM algorithm provided in this embodiment of the invention shows the curves of the change of the original residual and the dual residual with the number of iterations, verifying the convergence and computational efficiency of the alternating direction multiplier algorithm in the optimization problem of this invention. The results show that the algorithm can converge rapidly within a finite number of iterations, meeting the requirements of real-time market for decision-making speed, and supporting the advantages of this invention in terms of solution efficiency.

[0117] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A multi-timescale market bidding method for energy storage systems considering dynamic opportunity costs, characterized in that, Includes the following steps: S1: Acquire multi-source data from the electricity market and energy storage systems, including real-time electricity prices, grid load, energy storage state of charge (SOC), energy storage power, and renewable energy output data, and combine them with a pre-set prediction model to form a complete dataset containing the current state and future predictions. S2: Generative adversarial networks are used to simulate and generate various new energy output scenarios and their probabilities, and dynamic opportunity costs are evaluated based on discount theory; S3: By using the capacity state transfer equation and cross-market power balance constraints, the SOC and charging / discharging power of the energy storage system are dynamically correlated among the day-ahead market, intraday market, and real-time market. A multi-timescale market collaborative optimization model integrating dynamic opportunity cost constraints is constructed, and the evaluated dynamic opportunity cost is embedded as a constraint condition into the collaborative optimization model. S4: Using the Karush-Kuhn-Tucker condition and the alternating direction multiplier method, the collaborative optimization model is transformed into a distributed optimization problem, and the collaborative bidding strategies for the day-ahead market, intraday market and real-time market are obtained by solving the problem. S5: Perform security verification on the generated bidding strategy to ensure that it complies with the operating constraints of the energy storage system; S6: Executes strategies in a rolling time-domain manner and makes incremental optimizations and adjustments based on the latest market data and system status.

2. The multi-timescale market bidding method for energy storage systems considering dynamic opportunity costs according to claim 1, characterized in that, In step S2, the minimax game objective function of the generative adversarial network is defined as: ； in, As a generator, it will generate a random noise vector Mapped to output scenarios; It serves as a discriminator, used to determine the authenticity of the input scene; Data distribution contributing to the true history; After the Generative Adversarial Network (GAN) is trained, representative output scenarios are obtained by clustering and reducing the generated output scenarios. and the scene of output Corresponding probability of occurrence ,in .

3. The multi-timescale market bidding method for energy storage systems considering dynamic opportunity costs according to claim 2, characterized in that, In step S2, the dynamic opportunity cost is calculated using the following explicit quantification model: ； in, For time period Dynamic opportunity cost; It is a discount factor that reflects the time value of money; For time period The expected maximum return, through the output scenario and its probability of occurrence Perform calculations; For time period The actual achievable benefits; The explicit quantification model dynamically quantifies the impact of the current decision on future potential returns into cost constraints that can be embedded in the collaborative optimization model.

4. The multi-timescale market bidding method for energy storage systems considering dynamic opportunity costs according to claim 3, characterized in that, In step S3, the capacity state transfer equation includes: The equation for the state transmission from the daytime market to the intraday market: ,in For the current moment The initial electricity volume flowing from the previous day's market into the intraday market serves as the starting point for intraday market decisions. For the current moment Current market status of charge of energy storage; The discharge efficiency of the energy storage system; For the current moment The market's net power at present; This refers to the trading interval in the market prior to the current day; The state transfer equation from the intraday market to the real-time market: ,in For the current moment Initial electricity volume status flowing from the intraday market to the real-time market; The charging efficiency of the energy storage system; For the current moment Charging power in the market during the day; State evolution equation in real-time market: ,in for Real-time market status of energy storage charge; For the current moment Real-time market discharge power; To implement the market's trading time interval; The cross-market power balance constraint is defined as follows: ;in, , , These represent the energy storage system's performance in the day-ahead market, intraday market, and real-time market at the current moment. The declared net power.

5. The multi-timescale market bidding method for energy storage systems considering dynamic opportunity costs according to claim 4, characterized in that, In step S3, the objective function of the collaborative optimization model is a stochastic-robust hybrid form, defined as: ； in, For expected returns, In the scene The following benefits; Conditional risk value, , All are auxiliary variables; For robust margin, It is a set of uncertainties constructed based on Wasserstein distance, used to guard against errors in the probability distribution itself; Risk aversion coefficient; This is the robustness weighting coefficient.

6. The multi-timescale market bidding method for energy storage systems considering dynamic opportunity costs according to claim 5, characterized in that, In step S3, the collaborative optimization model adopts a three-layer nested structure, including: The upper-level problem aims to maximize the overall benefits of energy storage systems across multiple time scales by optimizing the pricing strategy parameters for each market. For mid-level problems, the dynamic opportunity cost is calculated in real time, and constraints are generated. Feedback to higher levels regarding the issue, including The preset opportunity cost threshold; The lower-level problem simulates the market clearing process with the goal of minimizing total social cost. The total social cost includes the power generation cost of conventional generator sets and the comprehensive cost of energy storage systems. The comprehensive cost of energy storage systems includes operating costs and dynamic opportunity costs.

7. A multi-timescale market bidding method for energy storage systems considering dynamic opportunity costs according to claim 6, characterized in that, The objective function, which aims to maximize the overall returns of energy storage systems across multiple time scales, is defined as follows: ； The constraints of the objective function, which aims to maximize the overall returns of energy storage systems across multiple time scales, are defined as follows: ; in, , , These are the pricing strategy parameters for energy storage systems in the day-ahead market, intraday market, and real-time market, respectively. , , These represent the maximum values ​​of the pricing strategy parameters for energy storage systems in the day-ahead market, intraday market, and real-time market, respectively. , , These refer to the daytime market, intraday market, and real-time market at the current moment. Electricity price; For the current moment The overall cost of energy storage systems; The objective function that aims to minimize total social cost is defined as follows: ； The constraints of the objective function aimed at minimizing total social cost include at least power balance constraints, generator output constraints, and energy storage operation constraints. The power balance constraint is defined as follows: ; The generator set output constraint is defined as follows: ; The energy storage operation constraints are defined as follows: ； in, For the first The power generation cost function of a generator set is used to describe its output. The relationship with cost can be expressed as: , , , All are unit characteristic coefficients; This is the overall cost function of the energy storage system; For the output of energy storage systems; For the first The output of the generator set; For the energy storage system at the current moment Net charge and discharge power, This indicates that the energy storage system is in a discharging state. This indicates that the energy storage system is in a charging state. This indicates that the energy storage system is in an idle state; Contribute to the total output of renewable resources; For nodes Load demand; , These represent the number of generator sets and load nodes, respectively. , The first The lower and upper limits of the output of the generator set; , These are the lower and upper limits of the SOC (State of Charge) for energy storage systems, respectively. , These represent the lower and upper limits of the energy storage system's output, respectively.

8. A multi-timescale market bidding method for energy storage systems considering dynamic opportunity costs according to claim 7, characterized in that, In step S4, the cooperative optimization model is transformed into a distributed optimization problem using the Karush-Kuhn-Tucker conditions and the alternating direction multiplier method, specifically as follows: The lower-level problem is transformed into a dual form using the Karush-Kuhn-Tucker condition, and the complementary relaxation condition is linearized using the Big M method, transforming the three-level nested collaborative optimization model into a single-level mixed integer linear programming (MILP) problem. The transformed single-level mixed integer linear programming (MILP) problem is solved in a distributed manner using the alternating direction multiplier method. Specifically, the global optimization variable is decomposed into subproblems corresponding to the day-ahead market, intraday market, and real-time market and solved in parallel. The dual variable is then iteratively updated until convergence is achieved through global consistency constraints.

9. A multi-timescale market bidding method for energy storage systems considering dynamic opportunity costs according to claim 8, characterized in that, In step S5, the safety verification includes at least hard constraint verification of the energy storage system's state of charge (SOC), charging / discharging power, and power change rate, expressed as: ； in, This represents the upper limit of the energy storage ramp-up rate.

10. A multi-timescale market bidding system for energy storage systems considering dynamic opportunity costs for implementing the method as described in any one of claims 1-9, characterized in that, include: The data acquisition layer is used to acquire or predict market electricity prices, grid load, renewable energy output, and energy storage system status data in real time. An optimized computing layer, connected to the data acquisition layer, includes a scene generation and classification unit and a core optimization solution engine; The scene generation and classification unit is used to generate uncertain scenes based on GAN; The core optimization engine embeds the dynamic opportunity cost quantification model and is used to execute the steps of the method described in any one of claims 1-9, to construct and solve the collaborative optimization model, and to generate a bidding strategy. The strategy execution layer, connected to the optimization calculation layer, includes a dynamic opportunity cost feedback module, a security verification module, and an energy storage device control terminal; The dynamic opportunity cost feedback module is used to feed the calculated dynamic opportunity cost value back to the optimization model; The security verification module is used to perform compliance checks on the generated bidding strategies; The control terminal of the energy storage device is used to convert the bidding strategy that has been verified for security into control commands and send them out for execution.

Images