Photovoltaic energy storage system capacity optimization configuration method considering user carbon income

Abstract

The invention relates to the technical field of power system automatic control, and discloses a photovoltaic energy storage system capacity optimization configuration method considering user carbon income, and the method comprises the steps: obtaining historical data, and generating a dynamic carbon value time sequence; based on nonlinear correlation and joint probability distribution among multi-dimensional uncertainty variables in generative adversarial network learning historical data, generating a scene set representing annual operation; constructing a two-stage stochastic optimization model taking expected annual net income maximization after risk adjustment as a target; solving a first-stage decision of the model to obtain optimal capacity configuration of the photovoltaic system and the energy storage system; and solving an optimal operation strategy of the photovoltaic energy storage system for each annual operation scene based on the determined capacity configuration. The system comprises a data acquisition module, a scene generation module, a model construction module and a model solving module. The invention aims to provide more economic and reliable photovoltaic energy storage system investment and operation decision support for industrial and commercial users.

Description

Technical Field

[0001] This invention relates to the field of power system automatic control technology, specifically to a method for optimizing the capacity configuration of a photovoltaic energy storage system that takes into account user carbon benefits. Background Technology

[0002] Against the backdrop of the ongoing global push for dual-carbon goals and the continuous reduction in the cost of energy storage technology, the application of distributed photovoltaic (PV) energy storage systems in the industrial and commercial sectors is rapidly expanding. These systems not only help industrial and commercial users reduce electricity costs and improve profitability, but also effectively increase the utilization efficiency of distributed PV. However, the deepening of current carbon reduction policies and the dynamic adjustment of electricity market pricing mechanisms pose significant challenges to the economic benefits that industrial and commercial users can obtain from configuring PV energy storage systems. Therefore, how to scientifically and rationally plan the capacity configuration of PV energy storage systems is of paramount importance for maximizing the economic benefits for industrial and commercial users, ensuring the reliability of system operation, and contributing to the achievement of dual-carbon goals.

[0003] While existing research on capacity optimization methods for photovoltaic (PV) energy storage systems has made some progress, it generally suffers from significant shortcomings in system economic evaluation. Most existing methods primarily focus on cost considerations, comprehensively taking into account factors such as PV output characteristics, user load fluctuations, and energy storage costs for capacity allocation, often neglecting the economic contribution of PV energy storage systems to industrial and commercial users in terms of carbon emissions. This singular cost perspective cannot fully reflect the true economic value of the system under the current policy environment, thus limiting the overall optimality of the configuration scheme. Therefore, there is an urgent need to research a capacity optimization method for PV energy storage systems that can fully calculate user carbon benefits, in order to address the shortcomings of existing technologies in economic benefit evaluation.

[0004] Therefore, this invention proposes a method for optimizing the capacity configuration of photovoltaic energy storage systems that takes into account user carbon benefits, in order to address the shortcomings of existing technologies. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a photovoltaic energy storage system capacity optimization configuration method that considers user carbon benefits. This method solves the problem that existing photovoltaic energy storage system capacity optimization configuration methods fail to fully consider multi-dimensional uncertainties when conducting economic evaluations, and the quantification of the carbon emission reduction economic value on the user side is not comprehensive enough, resulting in limited economic efficiency and robustness of the configuration scheme in actual operation.

[0006] To address the aforementioned technical problems, this invention provides a method and system for optimizing the capacity configuration of a photovoltaic energy storage system that takes into account user carbon benefits.

[0007] The first aspect of this invention provides a method for optimizing the capacity configuration of a photovoltaic energy storage system that takes into account user carbon revenue. This method determines the optimal capacity configuration and operation strategy of the photovoltaic energy storage system by constructing and solving a two-stage stochastic optimization model that takes into account risk factors.

[0008] In one embodiment, the method first acquires historical data of multidimensional uncertain random variables, including photovoltaic power output, user load, grid marginal carbon emission factor, carbon trading price, and carbon incentive response events. To accurately quantify the economic value of carbon emission reduction, the method generates a dynamic carbon value time series by multiplying the grid marginal carbon emission factor time series and the carbon trading price time series from the historical data point by point.

[0009] Subsequently, in order to characterize the coupling effect of the multidimensional uncertain random variables in the optimization model, a generative adversarial network is used to learn the variables in order to capture the nonlinear correlation and joint probability distribution among them, and a set of annual operating scenarios with corresponding occurrence probabilities are generated based on the learning results. This set of annual operating scenarios together constitutes a scenario set.

[0010] Based on the generated scenario set, a two-stage stochastic optimization model is constructed. The objective function of this model is set to maximize the risk-adjusted expected annualized net return. The calculation of the risk-adjusted expected annualized net return integrates the expected return and risk measurement for all annual operating scenarios. This calculation is based on the expected annualized net return of all annual operating scenarios, a user risk preference coefficient, and the conditional value of risk of the expected annualized net return at a preset confidence level, and is determined by the following formula for calculating the risk-adjusted expected annualized net return: ; In the formula, This represents the risk-adjusted expected annualized net return. The expected annualized net return for all annual operating scenarios; This refers to the user's risk preference coefficient. To achieve the expected annualized net return at a pre-set confidence level Conditional Value at Risk (VaR) measures the average net return in the worst-case scenario.

[0011] In one specific embodiment, the carbon emission reduction benefit included in the expected annualized net income consists of two parts: basic carbon income and incentive carbon income. The basic carbon income is calculated based on the dynamic carbon value time series generated in the preceding steps. The incentive carbon income is calculated based on the additional income obtained by the photovoltaic energy storage system in response to the carbon incentive response event.

[0012] To facilitate model solving, the calculation of the conditional value at risk is linearized by introducing auxiliary variables. Specifically, a first auxiliary variable and a set of second auxiliary variables corresponding one-to-one with the annual operating scenarios are introduced. The first auxiliary variable is used to characterize the distribution of net returns at a confidence level. The second auxiliary variable is used to quantify the difference between the actual net income and the said quantile value under each year's operating scenario.

[0013] The solution process for this two-stage stochastic optimization model consists of two stages. The first stage is the investment decision, and its solution is the installed capacity of the photovoltaic system, the rated power of the energy storage system, and the rated energy capacity of the energy storage system, determined after comprehensively weighing all annual operating scenarios. This decision is a unique investment decision that is valid for all annual operating scenarios.

[0014] The second stage of decision-making is operational decision-making. Under the capacity configuration conditions determined in the first stage, the optimal operational strategy for the photovoltaic energy storage system is solved for each annual operational scenario within the scenario set. The optimal operational strategy includes determining the energy storage charging and discharging power for each time period in each annual operational scenario, the power purchased from the grid, and a decision variable for deciding whether to participate in the carbon incentive response event. The determination of this optimal operational strategy is subject to a set of constraints, including a carbon incentive response constraint. This carbon incentive response constraint ensures that when the decision variable's state is "participate in the carbon incentive response event," the power purchased from the grid meets the emission reduction power requirements associated with that event.

[0015] A second aspect of the present invention provides a photovoltaic energy storage system capacity optimization configuration system that takes into account user carbon benefits. This system is used to perform the methods described in any of the foregoing embodiments. The system includes: The data acquisition module is used to acquire historical data including photovoltaic output, user load, grid marginal carbon emission factor, carbon trading price, and carbon incentive response events. The historical data together constitute a multidimensional uncertain random variable. The scenario generation module is used to generate a set of annual operating scenarios based on the multidimensional uncertain random variables. Each annual operating scenario is assigned an occurrence probability, and together they constitute a scenario set. The model building module is used to construct a two-stage stochastic optimization model based on the scenario set, with the objective function being to maximize the risk-adjusted expected annualized net return. The model solving module is used to solve the two-stage stochastic optimization model to determine the photovoltaic system installed capacity, energy storage system rated power and energy storage system rated energy capacity as the first-stage decision, and to obtain the optimal operation strategy of the photovoltaic energy storage system for each annual operation scenario in the scenario set as the second-stage decision.

[0016] This invention provides a method for optimizing the capacity configuration of a photovoltaic energy storage system while considering user carbon benefits. It offers the following advantages: 1. This invention employs a two-stage stochastic optimization model and generates an annual operating scenario set covering various uncertainties based on a generative adversarial network. This allows the invention to comprehensively capture the inherent correlation and long-term characteristics of photovoltaic power output, user load, and market price fluctuations. This enables the capacity configuration decision of photovoltaic energy storage systems to move beyond static or typical daily data and instead be based on a statistical trade-off of possible operating conditions throughout the year, thereby improving the economic efficiency and practical adaptability of the determined capacity configuration scheme in real-world applications.

[0017] 2. This invention integrates risk management into the optimization objective. By maximizing the risk-adjusted expected annualized net return, rather than simply the expected return, the optimization process can quantify and mitigate economic losses from potential extreme adverse scenarios. This consideration of risk ensures that the final investment decision, while pursuing high returns, is more robust. The configured photovoltaic energy storage system can more stably withstand the impact of market and environmental uncertainties, thus protecting the user's investment security.

[0018] 3. This invention delves deeper into the analysis and quantification of user carbon benefits. It not only accurately reflects the economic value of carbon emissions at different times by generating dynamic carbon value time series, but also distinguishes between basic and incentive benefits. This refined approach enables the system's capacity configuration to proactively optimize its utilization of its full economic potential in carbon reduction, particularly by acquiring additional value through responses to carbon incentive events, thus comprehensively enhancing the system's overall profitability.

[0019] 4. By optimizing the optimal operating strategy of the photovoltaic energy storage system, including participation in carbon incentive response events, in the second-stage decision-making process, this invention endows the configuration system with greater operational flexibility and grid friendliness. The system is no longer merely a passive unit for peak shaving and valley filling, but can proactively adjust its operating mode according to the grid's carbon emission reduction needs. This not only opens up new revenue channels for users but also enables the photovoltaic energy storage system to serve as a flexible regulatory resource, contributing to the low-carbon and stable operation of the power system. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is a system structure block diagram of the present invention; Figure 3 This is a schematic diagram of a typical daily photovoltaic power output curve according to the present invention.

[0021] The module consists of: 10. Data acquisition module; 20. Scene generation module; 30. Model building module; and 40. Model solving module. Detailed Implementation

[0022] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Reference Figure 1 , Figure 1 This is a flowchart illustrating a photovoltaic energy storage system capacity optimization configuration method considering user carbon revenue according to an embodiment of the present invention. The present invention provides a photovoltaic energy storage system capacity optimization configuration method considering user carbon revenue, which may include the following steps: S1. Obtain historical data, which includes photovoltaic output, user load, grid marginal carbon emission factor, carbon trading price, and carbon incentive response events. The historical data together constitute a set of multidimensional uncertain random variables.

[0024] S2 generates a set of annual operating scenarios based on a multidimensional set of uncertain random variables. Each annual operating scenario is assigned a probability of occurrence, and all annual operating scenarios together constitute a scenario set, which is used to represent the uncertainty of the future operating environment in the subsequent optimization model.

[0025] S3. Based on the scenario set, construct a two-stage stochastic optimization model. The objective function of the two-stage stochastic optimization model is set to maximize the risk-adjusted expected annualized net return. This objective function takes into account both the expected return level and the volatility risk of the return.

[0026] S4 solves the two-stage stochastic optimization model to obtain its first-stage decision. The first-stage decision is the investment decision, and its result is the installed capacity of the photovoltaic system, the rated power of the energy storage system, and the rated energy capacity of the energy storage system, which remain unchanged under all annual operating scenarios.

[0027] S5, under the capacity parameters of the photovoltaic system and energy storage system determined in S4, obtains the second-stage decision of the two-stage stochastic optimization model. The second-stage decision is the operational decision, and its result is the optimal operation strategy of the photovoltaic energy storage system for each annual operation scenario in the scenario set.

[0028] See attached document Figure 2 , Figure 2This is a structural block diagram of a photovoltaic energy storage system capacity optimization configuration system considering user carbon revenue, according to an embodiment of the present invention. The system is used to execute the aforementioned method and may include a data acquisition module 10, a scene generation module 20, a model building module 30, and a model solving module 40; wherein: The data acquisition module 10 is designed to acquire historical data including photovoltaic power output, user load, grid marginal carbon emission factor, carbon trading price, and carbon incentive response events, and output the historical data as a set of multidimensional uncertain random variables.

[0029] The scene generation module 20 is connected to the output end of the data acquisition module 10. Its function is to receive a set of multidimensional uncertain random variables and generate a set of annual operating scenarios with corresponding probabilities of occurrence, thus forming a scene set.

[0030] The model building module 30 is connected to the output of the scenario generation module 20. Its function is to build a two-stage stochastic optimization model with the objective function of maximizing the risk-adjusted expected annualized net return, based on the received scenario set.

[0031] The model solving module 40 is connected to the output of the model building module 30. Its function is to solve the constructed two-stage stochastic optimization model and output the photovoltaic system installed capacity, energy storage system rated power and energy storage system rated energy capacity as the first stage decision. At the same time, it outputs the photovoltaic energy storage system optimal operation strategy for each annual operation scenario as the second stage decision.

[0032] The following detailed description, with reference to the accompanying drawings, illustrates the specific implementation of the photovoltaic energy storage system capacity optimization configuration method considering user carbon revenue provided by the present invention.

[0033] Step S1: Acquire historical data and generate dynamic carbon value time series In one embodiment of the present invention, historical data acquisition and processing are first performed. This process is completed by the data acquisition module 10. The acquired historical data is a set of time-series data, with a time resolution of hourly or minute-level, to support subsequent refined modeling. Specifically, the acquired data dimensions include: photovoltaic power output data recorded in time series, user load data, grid marginal carbon emission factor data, and carbon trading price data. Simultaneously, historical carbon incentive response event data also needs to be acquired. This event data may include the start and end times of the event, its duration, and the required power reduction amount. These historical data, describing the internal and external environment of the system from different dimensions, together constitute a set of multidimensional random variables characterizing the uncertainty of system operation.

[0034] After obtaining the aforementioned basic data, to accurately quantify the economic value of carbon emission reduction achieved by photovoltaic energy storage systems through grid replacement at different times, a dynamic carbon value time series needs to be generated. This generation process involves mapping and multiplying the obtained grid marginal carbon emission factor time series with the carbon trading price time series point-by-point. Specifically, for each point in the first time series (e.g., ...), ... The marginal carbon emission factor of the power grid at time ) and the same time point in the second time series ( Multiply the carbon trading price at (time) to obtain the carbon trading price at that point in time (time). The dynamic carbon value (at any given time). This process is performed on all time points, ultimately resulting in a new time series with the same temporal resolution as the original data, namely the dynamic carbon value time series.

[0035] The generated dynamic carbon value time series will be passed to the model building module 30 and will serve as the core basis for calculating the basic carbon revenue component of the expected annualized net income when building the optimization model in step S3. In this way, the model can identify and utilize periods with high carbon value, thereby guiding the photovoltaic energy storage system to make more economical operating decisions.

[0036] Step S2: Generate the annual set of operating scenarios After acquiring and processing historical data, in order to effectively represent its uncertainty in the optimization model, this embodiment generates an annual set of operating scenarios. This process is completed by the scenario generation module 20. This module receives multidimensional uncertain random variables provided by the data acquisition module 10 and uses generative adversarial network (GAN) technology to learn and generate scenarios.

[0037] Generative Adversarial Networks (GANs) consist of two interconnected neural networks: a generator network and a discriminator network. The generator network takes a random noise vector as input and uses it to generate a synthetic data sample with the same structure as the original historical data, such as a complete annual data profile containing time series data on photovoltaic power output, user load, and dynamic carbon value.

[0038] The discriminator network functions as a binary classifier, trained to distinguish between real and synthetic data. During training, the discriminator network receives two types of input: real historical data samples from step S1 and synthetic data samples generated by the generator network. The discriminator's goal is to accurately determine whether each received data sample originates from a real or synthetic source.

[0039] The two networks undergo iterative optimization through an adversarial training process. In each iteration, the discriminator network updates based on its classification accuracy to improve its ability to distinguish between real and fake data. Simultaneously, the generator network updates based on feedback from the discriminator network (i.e., the degree to which its generated data is judged as fake), aiming to adjust its parameters to generate more realistic data, thus rendering the discriminator network unable to effectively distinguish between them.

[0040] This adversarial training process continues until the system reaches an equilibrium point. At this equilibrium point, the synthetic data generated by the generator network is highly consistent with the real historical data in terms of statistical properties and internal correlations, to the point that the discriminator network cannot distinguish between the two with an accuracy higher than random guessing. At this point, it can be considered that the generator network has successfully learned and captured the joint probability distribution of the original multidimensional uncertain random variables and their complex nonlinear correlations.

[0041] After training, a stable generator network is used to generate a large number of annual synthetic data profiles that conform to the learned distribution by inputting different random noise vectors. To reduce the computational complexity of subsequent model optimization, the generated profiles need to be reduced to form a representative and finite-sized scene set. One feasible reduction method is to use a clustering algorithm to group similar annual synthetic data profiles into one class, and select the most representative profile from each class as the representative scene of that class. The probability of occurrence of this representative scene is determined by the proportion of the original profiles contained in that class to the total number of generated profiles.

[0042] Finally, the scenario generation module 20 outputs a scenario set consisting of multiple representative annual operating scenarios. Each scenario in this set is assigned an occurrence probability, and the sum of the probabilities of all scenarios is 1. This scenario set is then passed to the model building module 30 as the basis for constructing a two-stage stochastic optimization model.

[0043] Step S3: Construct a two-stage stochastic optimization model After generating the annual operating scenario set in step S2, this embodiment performs the construction of a two-stage stochastic optimization model. This process is completed by the model building module 30, which establishes a mathematical optimization model that can accurately describe the system's economics and physical operating characteristics based on the generated scenario set.

[0044] In this embodiment, the fundamental goal of model construction is to determine an optimal capacity configuration scheme to maximize annualized net revenue. This applies to any annual operating scenario within the scenario set. Its annualized net income This can be expressed as the difference between total revenue and total cost in this scenario. Specifically, the annualized net revenue is obtained by subtracting the annualized investment cost, annual operation and maintenance cost, and annual battery replacement cost from the electricity cost reduction revenue and carbon emission reduction revenue. The calculation formula is as follows: ; The specific calculation methods for each benefit and cost are as follows: Electricity price reduction revenue consists of two parts: electricity price reduction revenue based on electricity consumption and basic electricity price reduction revenue. ; ; ; In the formula, This indicates the revenue reduction from annual electricity bills; To reduce revenue from basic electricity charges; The total number of scheduling cycles in a year; for The electricity purchase price from the grid at any given time; for User load power at any given time; for The amount of electricity purchased from the power grid at any given time; The length of a single scheduling time interval; The number of months in a year; This is the basic electricity fee per unit.

[0045] The carbon emission reduction benefits mainly consist of the basic carbon benefits obtained by replacing high-carbon-emission grid electricity with photovoltaic power generation and energy storage discharge: ; In the formula, Indicates the carbon trading price; This represents the carbon emission factor of the power grid. In this embodiment, the product term... This refers to the dynamic carbon value generated in step S1. In addition to the basic carbon revenue, the total carbon reduction revenue also includes incentive carbon revenue obtained in response to carbon incentive events.

[0046] The annualized investment cost of a photovoltaic energy storage system for industrial and commercial users is calculated using the following formula: ; In the formula: This represents the initial investment cost per unit of photovoltaic power. This indicates the installed capacity of photovoltaic power. , , , These represent the unit capacity cost, rated capacity, unit power cost, and rated power of energy storage, respectively. This is the capital recovery factor, used to annualize the initial investment cost.

[0047] The annual operation and maintenance cost of a photovoltaic energy storage system for industrial and commercial users is calculated using the following formula: ; ; ; In the formula: This indicates the annual operation and maintenance cost of photovoltaic equipment; This indicates the annual operating and maintenance cost of energy storage equipment; This indicates the unit capacity operation and maintenance cost of photovoltaic equipment; This indicates the unit capacity operation and maintenance cost of energy storage equipment.

[0048] The formula for calculating the annual cost of battery replacement is as follows: ; In the formula: This indicates the price per unit capacity of the battery. This indicates the number of battery replacements during the system's lifespan. Indicates the first Depreciation factor for each battery replacement.

[0049] Due to the uncertainty of the system's future operation, it involves multiple annual operating scenarios. Therefore, the optimization objective is not targeted at a single scenario, but rather seeks the optimal overall performance across all possible scenarios. Thus, the optimization objective is set as maximizing the expected annualized net return. Its calculation method is the annualized net return for all scenarios. Perform a probability-weighted average.

[0050] Furthermore, to mitigate potential economic losses from extremely unfavorable scenarios, this embodiment introduces a risk metric based on the expected return. Ultimately, the model's objective function is constructed to maximize the risk-adjusted expected annualized net return, in the following form: ; In the formula, This represents the risk-adjusted expected annualized net return. The expected annualized net return for all annual operating scenarios; This is the user's risk preference coefficient, used to adjust the decision-maker's trade-off between expected return and risk aversion. To achieve the expected annualized net return at a pre-set confidence level Conditional Value at Risk (VaR) measures the average net return in the worst-case scenario.

[0051] To handle conditional value of risk in the optimization model In this embodiment, a first auxiliary variable is introduced. and a set of operational scenarios for each year The second auxiliary variable in a one-to-one correspondence Linearization is then performed. It can be represented as: ; At the same time, the following constraints are added to the model: ; ; In the formula, the first auxiliary variable Characterizing the distribution of net income at a confidence level The quantile value below; the second auxiliary variable Used for quantification in a scenario Below that, net income is below the percentile value. The difference.

[0052] To ensure the reasonableness of the model output, a set of constraints needs to be imposed, specifically including: Electric power balance constraint, at any time The sum of the power generated and the power purchased within the system equals the sum of the power consumed by the load and the power charged by the energy storage.

[0053] ; In the formula: Indicates a typical day The output power of photovoltaics at any given time; Indicates energy storage on a typical day Discharge power at any given moment; Indicates energy storage on a typical day The charging power at any given time.

[0054] Photovoltaic output constraints: The actual output power of photovoltaic power cannot exceed its theoretical maximum power generation, and curtailment of photovoltaic power is permitted. ; ; In the formula: This indicates the power output per unit of photovoltaic installed capacity in the region; Indicates a typical day Real-time photovoltaic power output; Indicates a typical day The power of light discarded at any given moment.

[0055] Energy storage operation constraints: The operation of an energy storage system is limited by its rated power, rated capacity, and operating status.

[0056] ; ; ; ; ; In the formula, and These are binary variables of 0 and 1 representing the charging and discharging states of energy storage, respectively. for State of charge (SOC) that stores energy at all times; and These are the charging and discharging efficiencies of energy storage, respectively. and These represent the lower and upper limits of the permissible state of charge for energy storage, respectively.

[0057] Carbon incentive response constraints: If the system decides to participate in a carbon incentive response event, then during the response period, the power it purchases from the grid must meet the emission reduction requirements of the event. ; In the formula, The 0-1 decision variable for determining whether to participate in the carbon incentive response event; The set of time periods during which carbon incentive response events continue; Baseline load power before response; Minimum emission reduction power requirements associated with the event; It is a sufficiently large constant used to conditionalize the constraints.

[0058] Step S4: Solve the first-stage decision problem After constructing the two-stage stochastic optimization model in step S3, this embodiment proceeds to solve the first stage of the model. This process is completed by the model solving module 40, and its goal is to determine the optimal capacity configuration of the photovoltaic energy storage system.

[0059] In a two-stage stochastic optimization model, the decision variables are divided into two stages. The first stage decision variables are investment decision variables, characterized by the fact that decisions must be made before future uncertainties (i.e., which specific year's operating scenario will occur) are revealed. Once these decisions are made, they are usually fixed throughout the system's entire lifecycle.

[0060] In this embodiment, the decision variables for the first stage specifically include: the installed capacity of the photovoltaic system. Rated power of energy storage system and the rated energy capacity of the energy storage system These three variables together define the physical scale and capacity configuration of a photovoltaic energy storage system.

[0061] The model solving module 40 determines the optimal values ​​of the three investment decision variables by solving the optimization problem constructed in step S3, which aims to maximize the risk-adjusted expected annualized net return. Essentially, this solution process arrives at a unique investment decision that is well-adapted to all possibilities after comprehensively weighing the potential operating conditions and economic benefits of all annual operating scenarios in the scenario set generated in step S2.

[0062] The determined and The value of is not the optimal solution for a specific scenario (such as the most favorable or the most unfavorable scenario), but rather, in a statistical sense, the capacity allocation scheme that maximizes the risk-adjusted expected annualized net return across all possible future scenarios. This decision reflects a comprehensive consideration of future uncertainties, thereby ensuring the robustness of the investment plan.

[0063] The final output of this step is a set of defined values: the optimal photovoltaic installed capacity, rated energy storage power, and rated energy storage capacity. These values ​​will then be used as known parameters in the calculations of the next step.

[0064] Step S5: Obtain the second-stage decision After determining the unique investment decision variable in step S4, this embodiment then performs the second stage of model solving to obtain the optimal operating strategy for the photovoltaic energy storage system. This process is also completed by the model solving module 40. and This step involves determining the installed capacity of the photovoltaic system in the first phase of decision-making. Rated power of energy storage system and rated energy capacity of energy storage system This is performed under the condition that the parameters are known. At this point, the original two-stage stochastic optimization problem can be decomposed into a set of independent scenarios, each corresponding to a different annual operating scenario. The corresponding optimization subproblems.

[0065] Model solving module 40 is designed for each annual operating scenario in the scenario set. Each of these subproblems is then solved. The goal is to achieve the desired outcome in this specific scenario by utilizing a scheduling system within a given equipment capacity. Annualized net income Maximize. The solution process determines a set of second-stage decision variables that vary with the scenario.

[0066] This set of second-stage decision variables together constitutes the scenario. The optimal operating strategy for a photovoltaic energy storage system is as follows. This strategy is a detailed set of guidelines for the system's operation during various scheduling periods throughout the year. The instruction set for how to run. Specifically, the optimal running strategy includes the following decision variables: Each time period Power purchased from the grid ; Each time period Charging power of energy storage system With discharge power ; Each time period Binary variables corresponding to the energy storage charge and discharge states and ; And the binary decision variables that determine whether to participate in the carbon-incentive response event in this scenario. .

[0067] This solution process generates a corresponding optimal operating strategy for each annual operating scenario in the scenario set. Therefore, the final output of step S5 is a strategy set, which contains the optimal operating schemes pre-planned by the system to cope with all possible future operating scenarios under a given capacity configuration.

[0068] To further verify the effectiveness of the method and system provided by this invention, a specific application example will be used to illustrate this below.

[0069] In this example, step S1 is first executed to obtain the historical daily electricity consumption data of an industrial user and the photovoltaic output data of its region. Through data processing, such as using the K-means clustering algorithm, a typical daily electricity consumption curve and a corresponding typical daily photovoltaic output curve for the user are obtained, as shown in the attached figure. Figure 3 As shown.

[0070] Subsequently, power grid data and technical and economic parameters related to photovoltaic and energy storage equipment in the region were obtained. The average carbon emission factor of the power grid was set at 0.5703 t / MWh, and the carbon trading price was set at 100 yuan / t. Time-of-use electricity price data for the region are shown in Table 1.

[0071] Table 1: Time-of-use Electricity Pricing Table The key equipment parameters involved in the photovoltaic energy storage system are shown in Table 2.

[0072] Table 2: Parameter Table of Photovoltaic Energy Storage System After data preparation is complete, the aforementioned data and parameters are used as input and substituted into the two-stage stochastic optimization model constructed in step S3. Finally, by executing steps S4 and S5, the model solving module 40 solves the optimization model.

[0073] After the solution is completed, the model outputs the following first-stage decision results: the optimal installed capacity of the photovoltaic system is 690kW, and the optimal capacity configuration of the energy storage system is 356kWh. This result provides users with a clear basis for equipment investment and construction that balances economic efficiency and risk, and intuitively demonstrates the operability and technical effectiveness of this invention in practical application scenarios.

[0074] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for optimizing the capacity configuration of a photovoltaic energy storage system considering user carbon revenue, characterized in that, The method includes the following steps: S1. Obtain historical data, which includes photovoltaic output, user load, grid marginal carbon emission factor, carbon trading price, and carbon incentive response events. The historical data together constitute a multidimensional uncertain random variable. S2. Based on the multidimensional uncertain random variables, a set of annual operating scenarios are generated. Each annual operating scenario is assigned an occurrence probability, and together they constitute a scenario set. S3. Based on the scenario set, construct a two-stage stochastic optimization model. The objective function of the two-stage stochastic optimization model is to maximize the risk-adjusted expected annualized net return. S4. Solve the first-stage decision of the two-stage stochastic optimization model to determine the installed capacity of the photovoltaic system, the rated power of the energy storage system, and the rated energy capacity of the energy storage system. S5. Under the condition that the installed capacity of the photovoltaic system, the rated power of the energy storage system and the rated energy capacity of the energy storage system are determined, the second stage decision of the two-stage stochastic optimization model is obtained. The second stage decision is the optimal operation strategy of the photovoltaic energy storage system for each annual operation scenario in the scenario set.

2. The method for optimizing the capacity configuration of a photovoltaic energy storage system considering user carbon revenue as described in claim 1, characterized in that, In step S1, after obtaining the historical data, the method further includes: The time series of marginal carbon emission factors in the power grid and the time series of carbon trading prices in the historical data are multiplied point by point to generate a dynamic carbon value time series. The dynamic carbon value time series is used to calculate the carbon emission reduction benefits included in the risk-adjusted expected annualized net return when constructing a two-stage stochastic optimization model.

3. The method for optimizing the capacity configuration of a photovoltaic energy storage system considering user carbon revenue as described in claim 1, characterized in that, Step S2, the step of generating a set of annual operating scenarios based on the multidimensional uncertain random variables, includes: Generative adversarial networks are used to learn the multidimensional uncertain random variables in order to capture the nonlinear correlation between photovoltaic output, user load, grid marginal carbon emission factor, carbon trading price and carbon incentive response events, and the annual operation scenario is generated based on the learning results.

4. The method for optimizing the capacity configuration of a photovoltaic energy storage system considering user carbon revenue as described in claim 1, characterized in that, In step S3, based on the scenario set, a two-stage stochastic optimization model is constructed. The objective function of the two-stage stochastic optimization model is to maximize the risk-adjusted expected annualized net return. This step includes: The calculation of the risk-adjusted expected annualized net return combines the expected return and risk measurement for all annual operating scenarios. The expected annualized net return based on all annual operating scenarios, a user's risk preference coefficient, and the conditional value at risk of the expected annualized net return at a preset confidence level are determined using the following risk-adjusted expected annualized net return calculation formula: ； In the formula, This represents the risk-adjusted expected annualized net return. The expected annualized net return for all annual operating scenarios; This refers to the user's risk preference coefficient. To achieve the expected annualized net return at a pre-set confidence level Conditional Value at Risk (VaR).

5. The method for optimizing the capacity configuration of a photovoltaic energy storage system considering user carbon revenue according to claim 4, characterized in that, The carbon emission reduction revenue included in the expected annualized net income consists of two parts: basic carbon revenue and incentive carbon revenue. The calculation of the basic carbon revenue is based on a dynamic carbon value time series generated by multiplying the time series of the grid marginal carbon emission factor with the time series of carbon trading prices point by point. The incentive carbon revenue is the revenue obtained by the photovoltaic energy storage system in response to carbon incentive response events.

6. The method for optimizing the capacity configuration of a photovoltaic energy storage system considering user carbon revenue according to claim 4, characterized in that, To calculate the conditional value of risk, a first auxiliary variable and a set of second auxiliary variables corresponding one-to-one with the annual operating scenarios are introduced, thereby linearizing the calculation of the conditional value of risk. The first auxiliary variable represents the quantile value of the net income distribution; the second auxiliary variable is used to quantify the difference between the net income and the quantile value in each annual operating scenario.

7. The method for optimizing the capacity configuration of a photovoltaic energy storage system considering user carbon revenue as described in claim 1, characterized in that, Step S4, which involves solving the first-stage decision of the two-stage stochastic optimization model to determine the installed capacity of the photovoltaic system, the rated power of the energy storage system, and the rated energy capacity of the energy storage system, includes the following steps: The installed capacity of the photovoltaic system, the rated power of the energy storage system, and the rated energy capacity of the energy storage system are determined after comprehensively weighing the operating conditions of all annual operating scenarios, thus forming a unique investment decision that is effective for all annual operating scenarios.

8. The method for optimizing the capacity configuration of a photovoltaic energy storage system considering user carbon revenue according to claim 1, characterized in that, In step S5, the second-stage decision-making process for determining the optimal operating strategy of the photovoltaic energy storage system for each annual operating scenario within the scenario set includes the following steps: The optimal operating strategy for the photovoltaic energy storage system includes determining the energy storage charging and discharging power, the power purchased from the grid, and a decision variable for deciding whether to participate in the carbon incentive response event in each annual operating scenario.

9. A method for optimizing the capacity configuration of a photovoltaic energy storage system considering user carbon revenue, as described in claim 8, is characterized in that... The determination of the optimal operating strategy of the photovoltaic energy storage system is subject to a set of constraints, including a carbon incentive response constraint. The carbon incentive response constraint is used to ensure that when the decision variable is to participate in a carbon incentive response event, the power purchased from the grid meets the emission reduction power requirements associated with the carbon incentive response event.

10. A photovoltaic energy storage system capacity optimization configuration system considering user carbon revenue, applied to the method described in any one of claims 1-9, characterized in that, The system includes: The data acquisition module is used to acquire historical data including photovoltaic output, user load, grid marginal carbon emission factor, carbon trading price, and carbon incentive response events. The historical data together constitute a multidimensional uncertain random variable. The scenario generation module is used to generate a set of annual operating scenarios based on the multidimensional uncertain random variables. Each annual operating scenario is assigned an occurrence probability, and together they constitute a scenario set. The model building module is used to construct a two-stage stochastic optimization model based on the scenario set, with the objective function being to maximize the risk-adjusted expected annualized net return. The model solving module is used to solve the two-stage stochastic optimization model to determine the photovoltaic system installed capacity, energy storage system rated power and energy storage system rated energy capacity as the first-stage decision, and to obtain the optimal operation strategy of the photovoltaic energy storage system for each annual operation scenario in the scenario set as the second-stage decision.

Images