Wind and light output scene generation method, electronic equipment and computer program product

Abstract

The invention relates to the technical field of new energy power generation, and provides a wind and light output scene generation method, electronic equipment and a computer program product. Comprising the steps of performing data enhancement on historical wind and light output data based on a generative adversarial network, and generating an enhanced wind and light output scene set; dimension reduction processing is carried out on the enhanced wind and light output scene set through a variational cyclic auto-encoder model; performing clustering analysis on the data after dimension reduction processing based on a K-means clustering algorithm to obtain a wind power typical scene and a photovoltaic typical scene; and combining the wind power typical scene and the photovoltaic typical scene to generate a wind-solar combined output typical scene. Strong randomness and volatility of new energy power generation can be comprehensively considered, a wind-light combined output typical scene is obtained by fusing complex features in a wind-light output process, the problem of accurate quantification of wind-light uncertainty is solved, and the method has important significance in supporting new energy consumption and power grid dispatching operation.

Description

Technical Field

[0001] This application relates to the field of new energy power generation technology, specifically to a method for generating wind and solar power output scenarios, electronic equipment, and computer program products. Background Technology

[0002] The installed capacity of new energy sources, represented by wind and solar power, has grown rapidly, and the power grid generally exhibits a high proportion of new energy. This high proportion of new energy grid connection, coupled with the intermittent, random, time-varying, and low utilization hours characteristics of its power generation, brings a series of new changes and challenges to power system dispatch and operation. These are mainly manifested in the following ways: First, the power structure of the new power system will shift from being dominated by controllable and continuously outputting thermal power to being dominated by new energy power generation with strong randomness and weak controllability. The power system faces uncertainty in both the source and load sides, and new energy will gradually shoulder the responsibility of power balance. Therefore, the quantification of new energy uncertainty will directly affect the system's power balance and relate to the risk of power curtailment and shortage. Second, the strong randomness and volatility of new energy power generation make it difficult to accurately extract the complex characteristics of its output process and quantify its randomness, posing a significant challenge to modeling the uncertainty of power grid dispatch.

[0003] Taking all the above factors into account, how to integrate the complex characteristics of wind and solar power output, establish a quantitative model of wind and solar uncertainty, and generate typical scenarios is of great significance for supporting the consumption of new energy and the operation of power grid dispatch. Summary of the Invention

[0004] In view of this, the embodiments of this application provide a method for generating wind and solar power output scenarios, an electronic device, and a computer program product, which can integrate the complex characteristics of wind and solar power output, establish a quantitative model of wind and solar uncertainty, generate typical scenarios, solve the problem of accurate quantification of wind and solar uncertainty, and are of great significance for supporting the consumption of new energy and the operation of power grid dispatch.

[0005] The first aspect of this application provides a method for generating a wind and solar power output scene, including:

[0006] Data augmentation of historical landscape power output data is performed based on generative adversarial networks to generate an enhanced landscape power output scene set.

[0007] The enhanced wind and solar power output scene set is dimensionality reduced by using a variational cyclic autoencoder model.

[0008] Cluster analysis was performed on the dimensionality-reduced data based on the K-means clustering algorithm to obtain typical wind power scenarios and typical photovoltaic scenarios;

[0009] The typical wind power scenario and the typical photovoltaic scenario are combined to generate a typical wind-solar combined power output scenario.

[0010] A second aspect of this application provides a wind and solar power output scene generation device, comprising:

[0011] The scene generation module is used to perform data augmentation on historical landscape power output data based on generative adversarial networks, and generate an enhanced landscape power output scene set.

[0012] The dimensionality reduction module is used to perform dimensionality reduction processing on the enhanced wind and solar power output scene set through a variational cyclic autoencoder model;

[0013] The clustering analysis module is used to perform clustering analysis on the dimensionality-reduced data based on the K-means clustering algorithm to obtain typical wind power scenarios and typical photovoltaic scenarios.

[0014] The combination module is used to combine the typical wind power scenario and the typical photovoltaic scenario to generate a typical wind-solar combined output scenario.

[0015] A third aspect of this application provides an electronic device, including a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the electronic device enables the wind and solar power output scene generation method provided in the first aspect of this application.

[0016] A fourth aspect of this application provides a computer program product including a computer program that, when run, causes the method described in the first aspect of this application to be performed.

[0017] The wind and solar power output scenario generation method provided in the first aspect of this application involves augmenting historical wind and solar power output data using a generative adversarial network (GAN) to generate an augmented wind and solar power output scenario set; performing dimensionality reduction on the augmented wind and solar power output scenario set using a variational cyclic autoencoder (VAE) model; performing cluster analysis on the dimensionality-reduced data using a K-means clustering algorithm to obtain typical wind power scenarios and typical solar power scenarios; and combining the typical wind power scenarios and typical solar power scenarios to generate a combined wind and solar power output typical scenario. This method comprehensively considers the strong randomness and volatility of new energy power generation, and by integrating the complex characteristics of the wind and solar power output process, it obtains a combined wind and solar power output typical scenario, solving the problem of accurately quantifying the uncertainty of wind and solar power, and is of great significance for supporting the consumption of new energy and grid dispatching.

[0018] It is understood that the beneficial effects of the second to fourth aspects mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic flowchart of a method for generating a wind and solar power output scene according to an embodiment of this application;

[0021] Figure 2 This is a schematic flowchart of a method for generating a wind and solar power output scene according to another embodiment of this application;

[0022] Figure 3 This is a schematic diagram of the landscape power output scene set generated by WGAN in this application;

[0023] Figure 4 This is a schematic diagram illustrating the dimensionality reduction and clustering results of the generated scene set in this application;

[0024] Figure 5 This is a schematic diagram of a typical scene generated using the WGAN+VRAE method;

[0025] Figure 6 This is a typical scenario diagram generated using the KDE + K-means method;

[0026] Figure 7 This is a typical scenario diagram generated using the MC + K-means method;

[0027] Figure 8 This is a schematic diagram of a typical scene of the combined wind and solar power output generated by this application;

[0028] Figure 9 This is a schematic diagram of the structure of the wind and solar power output scene generation device provided in the embodiments of this application;

[0029] Figure 10 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0030] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0031] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0032] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0033] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0034] like Figure 1 As shown, the wind and solar power output scene generation method provided in this application embodiment includes the following steps S101 to S104:

[0035] Step S101: Perform data augmentation on historical landscape power output data based on generative adversarial network to generate an augmented landscape power output scene set;

[0036] Step S102: Dimensionality reduction of the enhanced wind and solar power output scene set is performed using a variational cyclic autoencoder model;

[0037] Step S103: Perform cluster analysis on the dimensionality-reduced data based on the K-means clustering algorithm to obtain typical wind power scenarios and typical photovoltaic scenarios;

[0038] Step S104: Combine typical wind power scenarios and typical photovoltaic scenarios to generate typical wind-solar combined output scenarios.

[0039] This application uses Wasserstein Generative Adversarial Network (WGAN) to perform data augmentation on historical landscape data. Based on the generated scenes, a scene reduction method based on VRAE and K-means is constructed to integrate data-driven and complex landscape feature information to obtain typical landscape scenes. Finally, the Cartesian product of the landscape scenes is calculated to extract the final landscape output scene set, which solves the problem of accurately quantifying the uncertainty of landscape.

[0040] In one embodiment, historical landscape power output data is augmented using a generative adversarial network to generate an augmented landscape power output scene set, including:

[0041] The historical data of wind and solar power generation at the hourly scale over a long-term time series of installed capacity are input into a pre-trained adversarial network model, which outputs an enhanced set of wind and solar power output scenarios.

[0042] In one embodiment, the training process of an adversarial network model includes:

[0043] The historical data of wind and solar long-term installed capacity factor at the hourly scale is used as the input dataset and divided into training and test sets.

[0044] A generative adversarial network (GAN) model is constructed, consisting of a generator and a discriminator. The generator takes Gaussian noise as input and its loss function is minimized. Generate wind and solar power output scene data; the discriminator maximizes the discriminator loss function. Distinguish between real data and generated data;

[0045] Using the input data, the objective function is minimized and maximized. The generator and discriminator are trained adversarially until the preset convergence condition is met, thus completing the training of the adversarial network model.

[0046] in, For those following a Gaussian distribution random variables Calculate the mathematical expectation; This indicates that it follows a Gaussian distribution. Noise data; Represents generator Generate data; This indicates that the discriminator is responsible for the generated data. The rating; To conform to the true distribution of wind and solar power output processes Sample values Calculate the mathematical expectation; For the discriminator to sample values The rating, The objective function is denoted as .

[0047] In one embodiment, using the input data, the objective function is minimized and maximized. The generator and discriminator are trained adversarially until a preset convergence condition is met, including:

[0048] Fixed generator parameters, based on and The discriminator is optimized using gradient descent to maximize the objective function. ;

[0049] Fixed discriminator parameters, based on The generator is optimized using gradient descent to minimize the objective function. ;

[0050] Repeat the above steps until a data distribution is generated that matches... The Wasserstein distance is less than a preset threshold.

[0051] In application, to obtain a scene set that can effectively characterize the stochastic characteristics of new energy output, a wind and solar scene generation method based on WGAN is adopted. The specific steps are as follows:

[0052] Step S201: Collect historical data of wind and solar long-term installed capacity factors at the hourly scale, construct the input set of the network model, and split the training set and the test set.

[0053] Step S202, Model Architecture Design:

[0054] Based on the actual process of the scenery's output, the day is divided into 24 time periods. For sampled values, This represents the actual distribution of wind and solar power output. This indicates that it follows a Gaussian distribution. Noise data.

[0055] generator The input data is noisy data. The output data is the generated data sample. The goal is to learn and generate a data distribution similar to real data, making the generated data as close as possible to the actual power output process of wind and solar power. Loss function It can be represented as:

[0056] ;

[0057] In the formula, Represents the expected distribution; Represents generator Generate data; Discriminator Output.

[0058] Discriminator The input data consists of the generated data and the actual power output process of the wind and solar power system; the output data is the probability value that the data is real. Discriminator The goal is to learn how to accurately distinguish between real and generated data. Loss function It can be represented as:

[0059] ;

[0060] In generative adversarial networks, by constructing a value function To establish a minimax game model and The confrontation process between them is shown in the following formula:

[0061] ;

[0062] Step S203: Based on the real dataset and the noisy data Generate a time series, fix the parameters of the generator G, input the generated data and the real data into the discriminator D respectively, calculate the loss of the discriminator, update the network parameters of the discriminator, and repeat the above steps until the Wasserstein distance meets the set requirements.

[0063] Step S204: Evaluate the generated data, adjust the model architecture and training hyperparameters based on the evaluation results until the generated data quality meets the requirements, and the training ends. Save the trained model and generate landscape scenes.

[0064] In one embodiment, the training process of a variational recurrent autoencoder model includes:

[0065] The enhanced set of wind and solar power output scenarios is used as input data, and the data is divided into a training set and a validation set.

[0066] A variational recurrent autoencoder model is constructed, which includes an encoder, latent variable sampling, and a decoder.

[0067] The encoder and decoder of the variational recurrent autoencoder model are trained using the input data until the preset convergence condition is met, thus completing the training of the variational recurrent autoencoder model.

[0068] In one embodiment, the encoder and decoder of a variational recurrent autoencoder model are trained using input data until a preset convergence condition is met, including:

[0069] The input data is reconstructed through an encoder, latent variable sampling, and a decoder to obtain the reconstructed data;

[0070] Calculate the reconstruction loss and KL divergence loss of the reconstructed data, and update the model parameters using the gradient descent algorithm until the reconstruction loss and KL divergence loss are less than the preset convergence threshold.

[0071] In the application, during the extraction of typical wind and solar power output scenes, a VRAE model was used to reduce the dimensionality of the generated wind and solar scenes in order to improve data processing capabilities and clustering effects. In the VRAE model, the encoder uses an RNN recurrent neural network, and the decoder still uses the decoder part of VAE. The trained model is used to reduce the dimensionality of the original data, and the resulting latent variables are used as input data. The specific steps are as follows:

[0072] Step S301: Using the wind and solar power output scene generated by WGAN as input data, divide it into training set and validation set;

[0073] Step S302, Model Architecture Design:

[0074] The main components of a VRAE model include an encoder, latent variable sampling, and a decoder. The encoder encodes the input time series data into latent variables; latent variable sampling refers to generating latent variables using the mean and variance parameters output by the encoder through reparameterization techniques; and the decoder decodes the latent variables back into the input data dimension.

[0075] Step S303, Model Training:

[0076] The input data is processed through an encoder, latent variable sampling, and decoder to obtain reconstructed data. The reconstruction loss and KL divergence loss are calculated, the gradient is calculated, and the model parameters are updated. The above steps are repeated until the loss converges.

[0077] Step S304: After training, the original data is dimensionality reduced using the trained model. The resulting latent variables are then used as input data for cluster analysis using the K-means algorithm in subsequent steps.

[0078] In one embodiment, cluster analysis is performed on the dimensionality-reduced data based on the K-means clustering algorithm to obtain typical wind power scenarios and typical photovoltaic scenarios, including:

[0079] Step S401: Randomly select K data points as initial cluster centers, where K is the number of clusters set in advance;

[0080] Step S402: Calculate the distance between each data point and each cluster center, and assign the data point to the cluster containing the nearest cluster center;

[0081] Step S403: Calculate the mean of all data points in each cluster, use the mean as the new cluster center, and return to the step of calculating the distance between each data point and each cluster center and assigning the data point to the cluster containing the nearest cluster center. Repeat steps S402 to S403 until the stopping condition is met.

[0082] Step S404: Use the decoder of the variational cyclic autoencoder model to restore the data and obtain typical wind power scenarios and typical photovoltaic scenarios.

[0083] In the application, the stopping condition in step S403 includes at least one of the following:

[0084] (a) The change in the coordinates of the cluster centers is less than a preset threshold;

[0085] (b) The cluster assignment results of the data points are completely consistent in two consecutive iterations;

[0086] (c) The preset maximum number of iterations is reached.

[0087] In one embodiment, typical wind power scenarios and typical photovoltaic scenarios are combined based on Cartesian product operations.

[0088] In one embodiment, this application provides a method for generating a wind and solar power output scene, such as... Figure 2 The steps described above are included in all of the embodiments.

[0089] To illustrate this method in practice, a specific project example will be used:

[0090] (1) Project background and parameter settings

[0091] The effectiveness of the proposed method is verified using the actual wind and solar power output process of a river basin in southwestern my country as an example, with a wind and solar installed capacity of 15046MW selected in the region. Based on the above engineering background, the method provided in the embodiments of this application is applied to verify its effectiveness and feasibility. The original wind and solar power output process is used as input data for WGAN to generate wind and solar power output scenarios. Then, scene reduction is performed through VRAE and K-means, and data-driven and complex wind and solar feature information are fused to obtain typical wind and solar scenarios. Finally, the set of combined wind and solar power output scenarios is extracted by Cartesian product calculation. All programs run on a personal computer equipped with an Intel Core i7-9750H CPU with a 2.6GHz CPU and 16.0MB of RAM.

[0092] (2) Results Analysis

[0093] To compare the reliability of the proposed method in generating typical scenes, Monte Carlo simulation and kernel density estimation methods were set as comparison methods for scene generation. The generated typical landscape scene set was used as input data, and peak shaving optimization scheduling was performed through deterministic and stochastic models, respectively. Figure 3This presentation showcases the output averages of 10,000 landscape scene sets generated using WGAN, corresponding to the five label categories. To improve computational efficiency, VRAE was used to reduce the dimensionality of the generated landscape output scenes, and K-means was used to cluster the dimensionality reduction results, thereby obtaining typical landscape output scenes. Figure 4 The results of dimensionality reduction and clustering of the generated scene set are presented. To verify the reliability of the scene generation method proposed, this paper also uses Monte Carlo simulation and kernel density estimation methods to generate scenes. Figure 5 , Figure 6 and Figure 7 Typical scenes generated using VRAE, KDE, and MC methods are presented respectively. The Spearman correlation coefficient, DTW distance, and Euclidean distance of the scenes generated by the three methods are calculated for comparison and analysis. See Table 1 for details.

[0094]

[0095] By comparing the Spearman correlation coefficients, it can be seen that the typical scenarios generated by the proposed method have larger Spearman correlation coefficients, indicating that the fluctuations of the generated scenarios are closer to the actual power output process. By comparing the DTW distances, it can be seen that the typical scenarios generated by the proposed method have smaller DTW distances, indicating that the temporal correlation of the generated scenarios is more consistent with the actual power output process. By comparing the Euclidean distances, it can be seen that the typical scenarios generated by the proposed method have smaller Euclidean distances, indicating that the generated scenarios are closer to the real process. Therefore, the typical scenarios generated by the proposed method have higher reliability and effectiveness. Figure 8 The image displayed is a combination of typical landscape scenes generated by VRAE.

[0096] Overall, this model can comprehensively consider the strong randomness and volatility of new energy power generation. By integrating the complex characteristics of wind and solar power output, it can obtain typical scenarios of combined wind and solar power output, which is of great significance for new energy consumption and grid dispatch operation.

[0097] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0098] This application also provides a wind and solar power output scene generation device for performing the steps in the above-described wind and solar power output scene generation method embodiments. The wind and solar power output scene generation device can be a virtual device within an electronic device, run by the processor of the electronic device, or it can be the electronic device itself.

[0099] like Figure 9As shown, the wind and solar power output scene generation device 100 provided in this application embodiment includes:

[0100] The scene generation module 101 is used to perform data augmentation on historical landscape power output data based on generative adversarial network, and generate an enhanced landscape power output scene set.

[0101] Dimensionality reduction module 102 is used to reduce the dimensionality of the enhanced wind and solar power output scene set through a variational cyclic autoencoder model;

[0102] The clustering analysis module 103 is used to perform clustering analysis on the dimensionality-reduced data based on the K-means clustering algorithm to obtain typical wind power scenarios and typical photovoltaic scenarios.

[0103] The combination module 104 is used to combine typical wind power scenarios and typical photovoltaic scenarios to generate typical wind and solar power combined output scenarios.

[0104] In applications, the modules in the wind and solar power generation device can be software program modules, or they can be implemented through different logic circuits integrated in the processor, or they can be implemented through multiple distributed processors.

[0105] like Figure 10 As shown, this application embodiment also provides an electronic device 200, including: at least one processor 201 ( Figure 10 The diagram shows only one processor, memory 202, and computer program 203 stored in memory 202 and executable on at least one processor 201. When processor 201 executes computer program 203, it implements the steps in the various method embodiments described above.

[0106] In applications, electronic devices may include, but are not limited to, processors and memory. Those skilled in the art will understand that... Figure 10 This is merely an example of an electronic device and does not constitute a limitation on the electronic device. It may include more or fewer components than shown, or a combination of certain components, or different components.

[0107] In applications, the processor can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0108] In applications, memory can be an internal storage unit of an electronic device in some embodiments, such as a hard drive or RAM. In other embodiments, memory can be an external storage device of the electronic device, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, memory can include both internal and external storage units of the electronic device. Memory is used to store operating systems, applications, bootloaders, data, and other programs, such as program code for computer programs. Memory can also be used to temporarily store data that has been output or will be output.

[0109] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0110] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0111] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps described in the various method embodiments above.

[0112] This application provides a computer program product, including a computer program, which, when run on an electronic device, enables the electronic device to perform the steps described in the various method embodiments above.

[0113] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. A computer-readable medium can include at least: any entity or device capable of carrying computer program code to a device / electronic device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0114] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0115] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0116] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0117] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0118] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for generating a landscape power output scene, characterized in that, include: Data augmentation of historical landscape power output data is performed based on generative adversarial networks to generate an enhanced landscape power output scene set. The enhanced wind and solar power output scene set is dimensionality reduced by using a variational cyclic autoencoder model. Cluster analysis was performed on the dimensionality-reduced data based on the K-means clustering algorithm to obtain typical wind power scenarios and typical photovoltaic scenarios; The typical wind power scenario and the typical photovoltaic scenario are combined to generate a typical wind-solar combined power output scenario.

2. The method for generating a wind and solar power output scene as described in claim 1, characterized in that, The process of augmenting historical landscape power output data using generative adversarial networks to generate an enhanced landscape power output scene set includes: The historical data of wind and solar power generation at the hourly scale over a long-term time series of installed capacity are input into a pre-trained adversarial network model, which outputs an enhanced set of wind and solar power output scenarios.

3. The method for generating a wind and solar power output scene as described in claim 2, characterized in that, The training process of the adversarial network model includes: The historical data of wind and solar long-term installed capacity factor at the hourly scale is used as the input dataset and divided into training and test sets. A generative adversarial network (GAN) model is constructed, comprising a generator and a discriminator; wherein the generator takes Gaussian distributed noise as input and minimizes the generator loss function. Generate wind and solar power output scene data; the discriminator maximizes the discriminator loss function. Distinguish between real data and generated data; Using the input data, the objective function is minimized and maximized. The generator and the discriminator are subjected to adversarial training until a preset convergence condition is reached, thus completing the training of the adversarial network model. in, For those following a Gaussian distribution random variables Calculate the mathematical expectation; This indicates that it follows a Gaussian distribution. Noise data; Represents generator Generate data; This indicates that the discriminator is responsible for the generated data. The rating; To conform to the true distribution of wind and solar power output processes Sample values Calculate the mathematical expectation; For the discriminator to sample values The rating, The objective function is denoted as .

4. The method for generating a wind and solar power output scene as described in claim 3, characterized in that, The input data is used to minimize and maximize the objective function. The generator and the discriminator are subjected to adversarial training until a preset convergence condition is met, including: Fixed generator parameters, based on and The discriminator is optimized using gradient descent to maximize the objective function. ; Fixed discriminator parameters, based on The generator is optimized using gradient descent to minimize the objective function. ; Repeat the above steps until a data distribution is generated that matches... The Wasserstein distance is less than a preset threshold.

5. The method for generating a wind and solar power output scene as described in claim 1, characterized in that, The training process of the variational recurrent autoencoder model includes: The enhanced wind and solar power output scene set is used as input data, and the data is divided into a training set and a validation set. A variational recurrent autoencoder model is constructed, which includes an encoder, latent variable sampling, and a decoder. The encoder and decoder of the variational recurrent autoencoder model are trained using the input data until a preset convergence condition is met, thus completing the training of the variational recurrent autoencoder model.

6. The method for generating a wind and solar power output scene as described in claim 5, characterized in that, The step of training the encoder and decoder of the variational recurrent autoencoder model using the input data until a preset convergence condition is met includes: The input data is reconstructed using the encoder, latent variable sampling, and decoder to obtain reconstructed data; Calculate the reconstruction loss and KL divergence loss of the reconstructed data, and update the model parameters using the gradient descent algorithm until the reconstruction loss and the KL divergence loss are less than a preset convergence threshold.

7. The method for generating a wind and solar power output scene as described in claim 1, characterized in that, The K-means clustering algorithm is used to perform cluster analysis on the dimensionality-reduced data to obtain typical wind power scenarios and typical photovoltaic scenarios, including: K data points are randomly selected as the initial cluster centers; Calculate the distance between each data point and each cluster center, and assign the data point to the cluster containing the nearest cluster center; Calculate the mean of all data points within each cluster, use the mean as the new cluster center, and return to the step of calculating the distance between each data point and each cluster center and assigning the data point to the cluster containing the nearest cluster center, until the stopping condition is met; The data is restored using the decoder of the variational cyclic autoencoder model to obtain typical wind power scenarios and typical photovoltaic scenarios.

8. The method for generating a wind and solar power output scene as described in claim 1, characterized in that, The typical wind power scenario and the typical photovoltaic scenario are combined based on the Cartesian product operation.

9. An electronic device, characterized in that, The device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, characterized in that, when the processor executes the computer program, the electronic device performs the method as described in any one of claims 1-8.

10. A computer program product, characterized in that, Includes a computer program, which, when run, causes the method as described in any one of claims 1-8 to be performed.

Images