Energy storage optimal configuration method, system and equipment considering new energy power generation uncertainty scene and medium

Abstract

The invention discloses an energy storage optimal configuration method, system and device considering a new energy power generation uncertainty scene, and a medium. The method comprises the steps: obtaining historical output data of a new energy power station, and carrying out the preprocessing; extracting a first feature index from the preprocessed data, and splicing the first feature index with the meteorological features to generate a potential code; game training is carried out based on a first learning algorithm, and fluctuation characteristics of new energy output uncertainty are learned; constructing a nonlinear mapping relation among the potential codes, the noise distribution and the new energy power generation scene, and generating the new energy power generation scene based on the nonlinear mapping relation; and constructing an energy storage capacity configuration model, and performing optimization solution on the new energy power generation scene to obtain an energy storage optimal configuration scheme. The method has the advantages of high adaptability, high feature interpretation, excellent optimization result reliability and the like, and is suitable for energy storage planning design and operation decision of a new energy power station and a regional power grid.

Description

Technical Field

[0001] This invention relates to the field of energy storage optimization configuration technology, and in particular to energy storage optimization configuration methods, systems, equipment and media that take into account the uncertainties of new energy power generation scenarios. Background Technology

[0002] Due to the influence of meteorological conditions and geographical environment, new energy power generation has significant randomness, intermittency and fluctuation characteristics, making it difficult to accurately predict the output. This poses challenges to the grid dispatch and operation, and also causes problems such as difficulty in new energy consumption and serious curtailment of wind and solar power, which seriously restricts the development of large-scale grid connection of new energy.

[0003] To improve the absorption of new energy sources and stabilize the operation of the power system, energy storage systems are widely considered an effective technological means. Energy storage can absorb excess electricity when there is a surplus in power generation and release energy when there is a shortage, achieving peak shaving, valley filling, frequency regulation, and voltage regulation. However, traditional energy storage configuration methods are usually based on static scenarios or a small amount of typical daily data, failing to fully consider the uncertain characteristics of new energy output, resulting in large deviations in energy storage capacity configuration and insufficient economic efficiency. Existing new energy scenario generation methods mostly rely on probabilistic statistical modeling or Monte Carlo sampling, such as using the distribution characteristics of historical data to build a probability model and then generating power generation scenarios through random sampling. Although this can describe the randomness of new energy output to a certain extent, it has two problems: firstly, it requires accurate modeling of the probability distribution, making it difficult to capture complex nonlinear characteristics; secondly, the generated scenarios lack diversity and cannot effectively express the temporal correlation and abrupt change characteristics of new energy output. In addition, in recent years, new energy output modeling methods based on machine learning and deep learning have gradually emerged. Some studies have attempted to use models such as generative adversarial networks and variational autoencoders to generate new energy scenarios, but they often suffer from poor feature interpretability and uncontrollable generated sample styles.

[0004] Therefore, there is an urgent need for a new energy scenario generation method that can take into account both feature interpretability and scenario diversity, so as to achieve coordinated planning and efficient utilization of new energy and energy storage systems. Summary of the Invention

[0005] In view of the aforementioned existing problems, the present invention is proposed.

[0006] Therefore, this invention provides an energy storage optimization configuration method, system, device, and medium that takes into account the uncertainty of new energy power generation scenarios, solving the problem that existing technologies are unable to maintain the interpretability and controllability of new energy output fluctuation characteristics while ensuring the diversity of generation scenarios.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides an energy storage optimization configuration method considering uncertainties in new energy power generation scenarios, including: Acquire historical power output data of new energy power plants and perform preprocessing; The first feature index is extracted from the preprocessed data, and the first feature index is concatenated with meteorological features to generate a potential code; Game theory training is conducted based on the first learning algorithm to learn the fluctuation characteristics of uncertainty in new energy output. A nonlinear mapping relationship is constructed between the potential coding, noise distribution, and new energy power generation scenario, and a new energy power generation scenario is generated based on the nonlinear mapping relationship; An energy storage capacity configuration model is constructed, and the optimal energy storage configuration scheme is obtained by optimizing the solution for the new energy power generation scenario.

[0008] As a preferred embodiment of the energy storage optimization configuration method considering uncertainties in new energy power generation scenarios described in this invention, the generation of potential codes includes: The first feature index is extracted from the preprocessed data, including the average daily power output, peak-to-valley difference, and maximum climbing power; daily temperature and humidity information are obtained based on historical meteorological data. The first feature index and meteorological features are concatenated as latent variables to form a latent code.

[0009] As a preferred embodiment of the energy storage optimization configuration method considering the uncertainty scenario of new energy power generation described in this invention, wherein: the step of learning the fluctuation characteristics of the uncertainty of new energy output based on the game training of the first learning algorithm includes: The latent code and random noise are used together as input to the generator; The generator outputs simulated power output scenarios and real historical power output scenarios are distinguished by a discriminator, and a classifier is introduced to reconstruct and predict the latent codes in the generated scenarios. A first joint loss function is constructed, and through the alternating optimization of the generator, discriminator and classifier, the uncertainty fluctuation characteristics of new energy output are deeply extracted.

[0010] As a preferred embodiment of the energy storage optimization configuration method considering uncertainties in new energy power generation scenarios described in this invention, the step of constructing the nonlinear mapping relationship between the potential coding, noise distribution, and new energy power generation scenarios includes: Based on the trained generator network, a deterministic mapping from the latent encoding and noise distribution to the new energy power generation scenario is established; By adjusting the numerical combination of the latent codes, the output fluctuation characteristics and meteorological correlation features of the generated scene can be controlled. By utilizing the aforementioned nonlinear mapping relationship, a set of new energy power generation scenarios with different power output styles is generated.

[0011] The beneficial effect of this preferred technical solution is that it can effectively characterize the fluctuation characteristics of wind power and photovoltaic power output without relying on complex probability modeling and sampling.

[0012] As a preferred embodiment of the energy storage optimization configuration method considering uncertainties in new energy power generation scenarios described in this invention, the step of acquiring historical output data of new energy power plants and performing preprocessing includes: Obtain historical power output data for new energy power plants, including predicted and actual values ​​for wind power and photovoltaic power; The historical output data of the aforementioned new energy power plant are normalized. Clean up missing or abnormal data, including discarding severely missing data and identifying and completing anomalies in relatively complete data.

[0013] As a preferred embodiment of the energy storage optimization configuration method considering uncertainties in new energy power generation scenarios described in this invention, the construction of the energy storage capacity configuration model includes: An objective function for an energy storage capacity configuration model is constructed with the optimization objectives of minimizing energy storage investment and construction costs and minimizing renewable energy curtailment costs. The energy storage investment and construction costs include energy storage capacity construction costs and energy storage power capacity construction costs, and the renewable energy curtailment costs include wind curtailment costs and solar curtailment costs.

[0014] The beneficial effects of this preferred technical solution are: improving the capacity for renewable energy absorption, reducing the curtailment rate of wind and solar power, and enhancing the economy and stability of the power system.

[0015] As a preferred embodiment of the energy storage optimization configuration method considering uncertainties in new energy power generation scenarios described in this invention, the constraints of the energy storage capacity configuration model include power balance constraints, energy storage operation constraints, conventional unit operation constraints, and new energy unit operation constraints.

[0016] Secondly, the present invention provides an energy storage optimization configuration system that considers uncertainties in new energy power generation scenarios, including: The data processing module is used to acquire historical power output data of new energy power plants and perform preprocessing. The feature extraction module is used to extract a first feature index from the preprocessed data and concatenate the first feature index with meteorological features to generate a latent code; The feature learning module is used for game-theoretic training based on the first learning algorithm to learn the fluctuation characteristics of uncertainty in new energy output. The mapping relationship construction module is used to construct a nonlinear mapping relationship between the potential coding, noise distribution and new energy power generation scenario, and generate a new energy power generation scenario based on the nonlinear mapping relationship; The optimization solution module is used to construct an energy storage capacity configuration model. By optimizing the solution for the new energy power generation scenario, the optimal energy storage configuration scheme is obtained.

[0017] Thirdly, the present invention provides an electronic device, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor executes the computer-executable instructions to implement the steps of an energy storage optimization configuration method considering uncertainties in new energy power generation scenarios.

[0018] Fourthly, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of an energy storage optimization configuration method considering uncertainties in new energy power generation scenarios.

[0019] Compared with existing technologies, the beneficial effects of this invention are as follows: Firstly, this invention constructs a historical power output dataset for new energy power plants, covering predicted and actual values ​​for wind and solar power. Data normalization and cleaning ensure the integrity and accuracy of the data. Interpretable feature indicators are extracted and used as latent coding inputs to characterize the uncertainty characteristics of new energy power generation. Based on generative adversarial networks, this invention performs deep learning on new energy output features, establishing a nonlinear mapping relationship between noise distribution and latent coding to new energy power generation scenarios, achieving diversified and interpretable new energy scenario generation. This generation model can effectively characterize the fluctuation characteristics of wind and solar power output without relying on complex probability modeling and sampling. Based on scenario generation, this invention constructs a capacity configuration model for energy storage systems with the optimization objectives of minimizing energy storage investment and construction costs and minimizing new energy curtailment costs, and introduces multiple constraints related to power balance, energy storage operation, conventional units, and new energy units. By optimizing and solving the generated new energy power generation scenarios, optimal energy storage configuration schemes adapted to different new energy uncertainty characteristics can be obtained, thereby improving the new energy absorption capacity, reducing wind and solar curtailment rates, and enhancing the economy and stability of the power system. The method of this invention has the advantages of strong adaptability, high feature interpretability, and high reliability of optimization results, and is applicable to the energy storage planning, design and operation decision-making of new energy power plants and regional power grids. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of the overall process logic of an energy storage optimization configuration method considering uncertainties in new energy power generation scenarios, provided as an embodiment of the present invention. Detailed Implementation

[0022] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0023] Example 1, referring to Figure 1 As one embodiment of the present invention, an energy storage optimization configuration method considering uncertainties in new energy power generation scenarios is provided, such as... Figure 1 The specific steps shown are as follows: S100: Acquire historical power output data of new energy power plants and perform preprocessing; S200: Extract the first feature index from the preprocessed data and concatenate the first feature index with meteorological features to generate a latent code; S300: Based on the first learning algorithm, game training is performed to learn the fluctuation characteristics of uncertainty in new energy output, construct the nonlinear mapping relationship between potential coding, noise distribution and new energy power generation scenario, and generate new energy power generation scenario based on the nonlinear mapping relationship; S400: Construct an energy storage capacity configuration model and obtain the optimal energy storage configuration scheme by optimizing and solving the new energy power generation scenario.

[0024] It should be noted that, to address the challenge of existing technologies in maintaining the interpretability and controllability of renewable energy output fluctuations while ensuring the diversity of generated scenarios, steps S100-S400 first construct a historical output dataset for renewable energy power plants, covering predicted and actual values ​​for wind and solar power. Data normalization and cleaning ensure data integrity and accuracy. Interpretable feature indicators are extracted and used as latent coding inputs to characterize the uncertainty of renewable energy generation. Based on generative adversarial networks, this invention performs deep learning on renewable energy output characteristics, establishing a nonlinear mapping relationship between noise distribution and latent coding to renewable energy generation scenarios, achieving diverse and interpretable renewable energy scenario generation. This generation model can effectively characterize the fluctuation characteristics of wind and solar power output without relying on complex probability modeling and sampling. Building upon scenario generation, this invention constructs a capacity configuration model for energy storage systems with the optimization objectives of minimizing energy storage investment and construction costs and minimizing renewable energy curtailment costs, introducing multiple constraints related to power balance, energy storage operation, conventional units, and renewable energy units. By optimizing the generated new energy power generation scenarios, optimal energy storage configuration schemes that adapt to the uncertainties of different new energy sources can be obtained, thereby improving the absorption capacity of new energy, reducing wind and solar curtailment rates, and enhancing the economy and stability of the power system. The method of this invention has advantages such as strong adaptability, high feature interpretability, and excellent reliability of optimization results, and is applicable to the energy storage planning, design, and operation decision-making of new energy power plants and regional power grids.

[0025] Example 2, based on the previous example, provides a specific implementation method for an energy storage optimization configuration method that considers the uncertainty of new energy power generation scenarios, to illustrate the technical means used in this method.

[0026] In this embodiment of the invention, step S100, which involves acquiring historical power output data of the new energy power plant and performing preprocessing, includes: Specifically, historical output data of new energy power plants are obtained, including predicted and actual values ​​of wind power and photovoltaic power, with a data sampling interval of 5 minutes, and stored in a CSV file; It should be noted that due to metering equipment malfunctions or data communication packet loss, the historical output data of new energy power plants may contain outliers and missing values. Some new energy power plants may have missing data for an entire day or even several consecutive days. Therefore, before constructing a historical output dataset for new energy power plants, the data needs to be preprocessed to ensure it provides complete and valid information. The preprocessing mainly includes data normalization and data cleaning.

[0027] Specifically, data normalization scales the original data to a small, specific range, preserving only the data's trend, thus facilitating comparative analysis of household electricity consumption data that were originally on different orders of magnitude. The zero-mean normalization method is used to normalize historical output data from renewable energy power plants, and its transformation function is as follows: in, For raw data The normalized value, and These represent the mean and standard deviation of the original data, respectively. The data after standardization with a mean of 0 has a mean of 0 and a standard deviation of 1, conforming to a standard normal distribution.

[0028] Specifically, historical output data of new energy power plants with serious deficiencies are discarded, while abnormal data is identified in relatively complete historical output data of new energy power plants, and outliers and missing values ​​are corrected and supplemented, thereby improving the completeness and accuracy of the data.

[0029] In an optional embodiment, the preprocessing process can also employ a sliding window-based local normalization method, which normalizes the data by calculating the local mean and standard deviation within the window, effectively eliminating the impact of intraday periodic fluctuations on the overall data distribution.

[0030] In another alternative embodiment, the preprocessing process can also apply data repair technology based on temporal similarity. By matching historical similar day patterns, the optimal alignment path is identified using a dynamic time warping algorithm, and similar segments are interpolated to fill in missing segments, thereby improving the temporal consistency and physical rationality of data repair.

[0031] In this embodiment of the invention, step S200, which involves extracting a first feature index from the preprocessed data and concatenating the first feature index with meteorological features to generate a potential code, includes: Specifically, the first feature index is extracted from the preprocessed data, including the average daily power output, peak-to-valley difference, and maximum ramp power; daily temperature and humidity information are obtained based on historical meteorological data. The first feature index and meteorological features are concatenated as latent variables to form a latent code c, which is beneficial for subsequent control of the style generated in the new energy power generation scenario.

[0032] Specifically, the average power output of new energy sources can reflect the average power generation level of new energy power plants in this scenario, and the calculation formula is as follows: in, The average power output of new energy sources Let T be the output of the new energy power station at the t-th time segment, where T is the total number of time segments in a day.

[0033] Specifically, the peak-valley difference in renewable energy output reflects the peak-valley difference in power output of renewable energy power plants in this scenario. The calculation formula is as follows: in, Peak-valley difference in energy output for new energy sources The maximum power output for new energy sources The minimum power output for new energy sources.

[0034] Specifically, the maximum ramp-up power of renewable energy can reflect the maximum power fluctuation level of renewable energy power plants within adjacent time periods under this scenario. The calculation formula is as follows: in, The maximum ramping power for renewable energy sources This refers to the power output of the new energy power plant at the (t-1)th time segment.

[0035] In an optional embodiment, the first characteristic indicator may also be the daily power output fluctuation frequency, which quantifies the frequent fluctuation characteristics of new energy power output by counting the number of times the power change rate exceeds a set threshold per unit time.

[0036] In another optional embodiment, the first feature index may also be the similarity of the daily power output curve shape, which uses dynamic time warping distance or curve principal component analysis score to characterize the similarity between the daily power output curve and the typical power output mode, so as to capture the power output time sequence shape features.

[0037] In this embodiment of the invention, step S300 includes the following sub-steps A1 and A2: In A1: Game theory training is performed based on the first learning algorithm to learn the fluctuation characteristics of uncertainty in new energy output; detailed steps include: The latent code and random noise are used together as input to the generator; The discriminator distinguishes between the simulated output scene and the real historical output scene output by the generator, and a classifier is introduced to reconstruct and predict the latent encoding in the generated scene. A first joint loss function is constructed, and through alternating optimization of the generator, discriminator and classifier, the uncertainty fluctuation characteristics of new energy output are deeply extracted.

[0038] In A2: A nonlinear mapping relationship is constructed between latent coding, noise distribution, and new energy power generation scenarios, and new energy power generation scenarios are generated based on this nonlinear mapping relationship; detailed steps include: Based on the trained generator network, a deterministic mapping from latent encoding and noise distribution to new energy power generation scenarios is established. By adjusting the numerical combination of latent codes, the output fluctuation characteristics and meteorological correlation features of the generated scene can be controlled. By utilizing nonlinear mapping relationships, a set of new energy power generation scenarios with different output styles is generated.

[0039] In an optional embodiment, the first learning algorithm can also be a generative model based on a conditional variational autoencoder, which learns the probability distribution of historical power output data through the encoder and introduces latent variable constraints such as meteorological conditions into the latent space, so that the decoder can generate new energy power output scenarios that meet specific conditions.

[0040] In another alternative embodiment, the first learning algorithm may also employ a standardized flow model, which gradually maps a simple distribution to a complex new energy output distribution through a series of reversible transformations, thereby achieving accurate modeling and controllable generation of output fluctuation characteristics.

[0041] In this embodiment of the invention, the fluctuation characteristics of the uncertainty in the output of new energy sources are learned based on information generation adversarial networks.

[0042] It's important to note that in traditional Generative Adversarial Networks (GANs), the generator has virtually no constraints. Its input is random, noisy data, and this noise is intertwined with useful features, resulting in overly free-flowing generated data that cannot produce scene data with a specific style. During the iterative training of a GAN, although the generator can learn the distribution of real data, the learned features are encoded in a complex and disordered manner. If these features can be separated and encoded, the data generated by the generator will be more diverse and interpretable.

[0043] Specifically, InfoGAN combines the original GAN ​​with feature learning by introducing mutual information, thus enabling the description and control of feature information without relying on external information or manually adding data labels. It mainly consists of a generator, a discriminator, and a classifier. In the generator's input, the random noise data z is divided into two parts: a latent code c composed of several interpretable latent variables and incompressible noise z'. The classifier then predicts the generator's output to obtain the predicted latent code c'.

[0044] Specifically, in information theory, mutual information is used to represent the amount of information between two variables. Assume the uncertainty of variable X is... Given variable Y, its uncertainty is: Then the reduction in uncertainty of variable X can be expressed as: Specifically, InfoGAN introduces mutual information. The data samples generated by characterization The degree of association between the latent code c and the generator is used to construct the constraint relationship between the latent code c and the generator. To make... The relationship with c is closer, requiring the maximization of mutual information. If the value of is , then the loss function becomes: in, This is the regularization parameter, which is typically set to 1.

[0045] It should be noted that, due to mutual information There are marginal probabilities that are difficult to calculate; variational inference is used to construct an approximate distribution. Replacing it with this will help reduce computational complexity: The final loss function of InfoGAN is as follows: In an optional embodiment, the construction of the first joint loss function can also introduce feature matching loss on top of adversarial loss, thereby improving training stability and the diversity of generated scenes by minimizing the difference in feature statistics between the generated scene and the real scene in the intermediate layer of the discriminator.

[0046] In another alternative embodiment, the construction of the first joint loss function can also combine Wasserstein distance and gradient penalty term to construct adversarial loss, and add latent encoding reconstruction loss to strengthen the semantic correlation between the generated scene and the conditional encoding by reconstructing the input latent variables through the encoder network.

[0047] In this embodiment of the invention, fully mining the implicit relationships in massive historical data through historical output datasets specifically includes the following steps: Assume there is N Complete historical observation data of new energy power plants ,in , indicating the length of the time series; This represents the number of new energy power plants. Scene generation utilizes historical data on new energy power generation as a training set to build and train a generative model. By learning from historical data, the implicit relationships within the massive historical data are fully explored, enabling the generative model to generate data... The distribution patterns of the generated data align with those of historical training data, effectively characterizing the uncertainties in renewable energy generation. These generated time-series data constitute the generated new energy generation scenarios, satisfying the following objective function: in, For probability operators, This is the set probability distribution error.

[0048] In this embodiment of the invention, constructing the nonlinear mapping relationship between noise distribution and potential coding and new energy power generation scenarios specifically includes the following steps: For the generator, the input is a set of random noise data. z And the latent code composed of several interpretable latent variables. c ,by express z The probability distribution, and at the same time This model represents the probability distribution of historical scene data, and the output is the generated new energy power generation scene. During training, the generator and discriminator continuously learn and optimize, updating their parameters to improve their generation and discrimination capabilities respectively, in order to find a Nash equilibrium between them. As training converges, the generator network eventually captures the latent distribution of real data. The trained new energy power generation scene generation model can generate controllable and diverse new energy power generation scenes by inputting noise distribution and manually controlled latent encoding.

[0049] It should be noted that step S300 above utilizes information generative adversarial networks to establish a nonlinear mapping from potential codes and noise to power generation scenarios through game training. This enables the efficient generation of diverse and interpretable new energy output scenarios without the need for complex probabilistic assumptions, significantly improving the realism and coverage of scenario generation.

[0050] In this embodiment of the invention, step S400 above constructs an energy storage capacity configuration model, and by optimizing and solving the new energy power generation scenario, obtains the optimal energy storage configuration scheme, including: Specifically, the objective function of the energy storage capacity configuration model is constructed with the optimization goals of minimizing energy storage investment and construction costs and minimizing renewable energy curtailment costs. Here, energy storage investment and construction costs include both energy storage capacity construction costs and energy storage power capacity construction costs, while renewable energy curtailment costs include wind curtailment costs and solar curtailment costs. The formula is as follows: in, For energy storage investment and construction costs; Costs of curtailing renewable energy; This refers to the set of areas where new energy transmission is blocked. For areas where new energy transmission is blocked i The energy storage capacity to be constructed; For areas where new energy transmission is blocked iThe energy storage capacity under construction; For areas where new energy transmission is blocked i Construction cost per unit energy capacity of energy storage; For areas where new energy transmission is blocked i Construction cost per unit power capacity of energy storage; The discount rate; For areas where new energy transmission is blocked i The service life of energy storage; For areas where new energy transmission is blocked i medium wind farm exist t The amount of wind curtailed during a given time period; Cost per unit of wind curtailment; For areas where new energy transmission is blocked i China Photovoltaic Power Station exist t The amount of light wasted during a given time period; Cost per unit of abandoned light.

[0051] Specifically, the constraints of the energy storage capacity configuration model include power balance constraints, energy storage operation constraints, conventional unit operation constraints, and new energy unit operation constraints.

[0052] Specifically, the formula for the power balance constraint is expressed as follows: in, , The first t At any given time, the transmission of new energy is blocked. i The discharge power and charging power of energy storage.

[0053] Specifically, the formula for energy storage operation constraints is expressed as follows: in, For areas where new energy transmission is blocked i Energy storage in the first t Energy state at each scheduling moment; This method uses a 1-hour time window as a scheduling time window. , To improve the charging and discharging efficiency of energy storage; For areas where new energy transmission is blocked i Energy storage in the first tThe charging 0-1 variable is set at each scheduling time. It is 1 when the energy storage is in the charging state, and 0 otherwise. For areas where new energy transmission is blocked i Energy storage in the first t The discharge 0-1 variable at each scheduling moment is set to 1 when the energy storage is in the discharge state, and 0 otherwise.

[0054] Specifically, the formula for the operating constraints of conventional generating units is expressed as follows: in, , These are the areas where new energy transmission is blocked. i The minimum and maximum generating power of thermal power unit g; , These are the areas where new energy transmission is blocked. i The downhill and uphill gradient rates of thermal power unit g; , For areas where new energy transmission is blocked i Thermal power unit g in the first t The number of time-series segments for continuous power-on and power-off at each scheduling moment. For areas where new energy transmission is blocked i The minimum number of sequential segments for continuous start-up and shutdown of thermal power unit g.

[0055] Specifically, the formula for the operating constraints of new energy generating units is expressed as follows: in, , These are the upper limits for wind power and solar power output, respectively.

[0056] It should be noted that step S400 above constructs an energy storage configuration model with optimal economic efficiency as its goal, and comprehensively considers multiple constraints of system operation to achieve energy storage capacity optimization decision-making under uncertain scenarios. It can generate customized energy storage configuration schemes for different new energy fluctuation characteristics, effectively improving the system's new energy absorption capacity and overall operational economy.

[0057] Example 3: This example provides an energy storage optimization configuration system that considers uncertainties in new energy power generation scenarios, including: The data processing module is used to acquire historical power output data of new energy power plants and perform preprocessing. The feature extraction module is used to extract the first feature index from the preprocessed data and concatenate the first feature index with meteorological features to generate a latent code; The feature learning module is used for game-theoretic training based on the first learning algorithm to learn the fluctuation characteristics of uncertainty in new energy output. The mapping relationship construction module is used to construct a nonlinear mapping relationship between potential coding, noise distribution and new energy power generation scenarios, and generate new energy power generation scenarios based on the nonlinear mapping relationship; The optimization solution module is used to build an energy storage capacity configuration model. By optimizing the solution for new energy power generation scenarios, the optimal energy storage configuration scheme is obtained.

[0058] It should be noted that the technical solution of the energy storage optimization configuration system considering the uncertainty scenario of new energy power generation is based on the same concept as the technical solution of the energy storage optimization configuration method considering the uncertainty scenario of new energy power generation described above. For details not described in detail in the technical solution of the energy storage optimization configuration system considering the uncertainty scenario of new energy power generation in this embodiment, please refer to the description of the technical solution of the energy storage optimization configuration method considering the uncertainty scenario of new energy power generation described above.

[0059] The above-mentioned unit modules can be embedded in the processor of the electronic device in hardware form or independent of it, or they can be stored in the memory of the electronic device in software form, so that the processor can call and execute the corresponding operations of the above modules.

[0060] This embodiment also provides an electronic device, which includes a processor, a memory, a communication interface, a display screen, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements an energy storage optimization configuration method considering the uncertainties of new energy power generation scenarios. The display screen can be a liquid crystal display (LCD) or an e-ink display. The input device can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the device's casing, or an external keyboard, touchpad, or mouse.

[0061] This embodiment also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method proposed in the above embodiments.

[0062] The storage medium proposed in this embodiment belongs to the same inventive concept as the method proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0063] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory, random access memory, flash memory, hard disk, or optical disk, and includes several instructions to cause an electronic device (which may be a personal computer, server, or network device, etc.) to execute the method of the embodiments of the present invention.

[0064] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for optimizing configuration of energy storage considering uncertainty scenarios of new energy power generation, characterized in that, The method comprises the following steps: acquiring historical output data of a new energy power station and preprocessing the data; extracting first feature indicators from the preprocessed data, and splicing the first feature indicators with meteorological features to generate latent encoding; conducting game training based on a first learning algorithm to learn fluctuation characteristics of new energy output uncertainty; constructing a nonlinear mapping relationship between the latent encoding, noise distribution and new energy generation scenarios, and generating new energy generation scenarios based on the nonlinear mapping relationship; constructing a storage capacity configuration model, and obtaining an optimal storage configuration scheme by optimizing and solving the new energy generation scenarios. 2.The method of claim 1, wherein, The method for generating latent encoding comprises the following steps: extracting first feature indicators from the preprocessed data, including average power, peak-valley difference and maximum climbing power of daily output; and obtaining temperature and humidity information of each day according to historical meteorological data; splicing the first feature indicators and meteorological features as hidden variables to form latent encoding. 3.The method of claim 2, wherein, The method for conducting game training based on a first learning algorithm to learn fluctuation characteristics of new energy output uncertainty comprises the following steps: taking the latent encoding and random noise as inputs of a generator; distinguishing simulated output scenarios output by the generator from real historical output scenarios by a discriminator, introducing a classifier to reconstruct and predict the latent encoding in the generated scenarios; constructing a first joint loss function, and realizing deep extraction of fluctuation characteristics of new energy output uncertainty by alternating optimization of the generator, the discriminator and the classifier. 4.The method of claim 3, wherein, The method for constructing a nonlinear mapping relationship between the latent encoding, noise distribution and new energy generation scenarios comprises the following steps: establishing a deterministic mapping from the latent encoding and noise distribution to new energy generation scenarios based on a trained generator network; controlling output fluctuation characteristics and meteorological correlation characteristics of generated scenarios by adjusting numerical combinations of the latent encoding; generating a set of new energy generation scenarios with different output styles by using the nonlinear mapping relationship. 5.The method of claim 1, wherein, The method for acquiring historical output data of a new energy power station and preprocessing the data comprises the following steps: acquiring historical output data of a new energy power station, including predicted values and actual values of wind power and photovoltaic power; normalizing the historical output data of the new energy power station; cleaning data with missing or abnormal values, including discarding data with serious missing values, and identifying and completing relatively complete data. 6.The method of claim 5, wherein, The method for constructing a storage capacity configuration model comprises the following steps: constructing an objective function of the storage capacity configuration model with the lowest storage investment and construction cost and the lowest new energy curtailment cost as optimization objectives, wherein the storage investment and construction cost includes storage energy capacity construction cost and storage power capacity construction cost, and the new energy curtailment cost includes wind curtailment cost and light curtailment cost. 7.The method of claim 6, wherein, The constraint conditions of the storage capacity configuration model include power balance constraint, storage operation constraint, conventional unit operation constraint and new energy unit operation constraint.

8. The energy storage optimal configuration system considering the uncertainty scenario of new energy power generation, applying the energy storage optimal configuration method considering the uncertainty scenario of new energy power generation according to any one of claims 1-7, characterized in that, The method comprises the following steps: a data processing module is configured to acquire historical output data of a new energy power station and preprocess the data; The feature extraction module is configured to extract a first feature index from the preprocessed data, splice the first feature index with meteorological features, and generate latent encoding; The feature learning module is configured to learn fluctuation characteristics of new energy output uncertainty based on a first learning algorithm for game training. The mapping relationship construction module is configured to construct a nonlinear mapping relationship between the latent encoding, noise distribution, and new energy generation scenarios, and generate new energy generation scenarios based on the nonlinear mapping relationship. The optimization solving module is configured to construct an energy storage capacity configuration model, and obtain an optimal energy storage configuration scheme by optimizing and solving the new energy generation scenarios. 9.An electronic device comprising a memory and a processor, the electronic device characterized by: The memory is configured to store computer executable instructions, and the processor executes the computer executable instructions to implement the steps of the energy storage optimization configuration method considering new energy generation uncertainty scenarios according to any one of claims 1-7.

10. A computer-readable storage medium having computer-executable instructions stored thereon, characterized in that: The computer executable instructions are executed by the processor to implement the steps of the energy storage optimization configuration method considering new energy generation uncertainty scenarios according to any one of claims 1-7.

Images