Source-network-load combined scene generation method and system considering meteorological coupling factors

Abstract

The invention discloses a source-network-load combined scene generation method and system considering meteorological coupling factors, and the method comprises the steps: carrying out the correlation analysis of historical meteorological data and historical source-network-load data, obtaining key meteorological factors, and carrying out the clustering and grouping of the key meteorological factors; according to the clustering grouping result of the key meteorological factors, dividing the source-network-load data corresponding to the time to which the key meteorological factors belong into groups, and obtaining a source-network-load data grouping result; and constructing and training a generative adversarial network model of each group according to a source-network-load data grouping result, inputting the source-network-load data into the trained generative adversarial network model, and generating a source-network-load joint scene set. According to the method, meteorological factors are used as a key climatic factor set of source-network-load correlation, a complex scene reduction link required by scene random matching after the source, the network and the load are respectively generated is not needed, and the scene generation efficiency is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of smart distribution networks, and in particular to a method and system for generating a source-grid-load joint scenario taking into account meteorological coupling factors. Background Art

[0002] Constructing uncertainty scenario models for renewable energy, transmission lines and loads is the key to achieving safe and stable operation, economic dispatch, and reasonable planning of power systems. Most of the existing scenario generation methods are probability model methods and classic scenario methods. The probability model method generates output or load scenarios based on statistical assumptions or probability distributions, combined with sampling methods such as Monte Carlo. These statistical assumptions may not hold true or be difficult to verify in practice, and relying on certain probability predictions may limit the diversity of generated scenarios; the classic scenario generation method uses data mining and scenario reduction techniques to reduce or optimize large-scale historical scenarios to generate a set of classic scenarios, which requires cumbersome processing and the generated scenarios are difficult to be different from historical data.

[0003] In addition to generating output, transmission lines, loads and other scenarios separately, in order to meet the needs of scheduling and planning, the correlation between sources, networks and loads needs to be considered when generating scenarios. At present, most of them only consider the correlation between power generation side output such as wind and light, while ignoring the correlation between sources, networks and loads. Summary of the invention

[0004] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0005] In view of the above existing problems, the present invention is proposed. Therefore, the present invention provides a source-grid-load joint scenario generation method taking into account meteorological coupling factors, which solves the problem of how to use meteorological factors as coupling features to construct a typical "source-grid-load" joint scenario set under various climate conditions without explicit models and probability distributions.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a source-grid-load joint scenario generation method taking into account meteorological coupling factors, comprising:

[0008] Obtain historical meteorological data and historical source-grid-load data;

[0009] Performing correlation analysis on the historical meteorological data and historical source-grid-load data to obtain key meteorological factors, and clustering and grouping the key meteorological factors;

[0010] According to the clustering grouping results of key meteorological factors, the source-grid-load data corresponding to the time of the key meteorological factors are divided into groups to obtain the source-grid-load data grouping results;

[0011] According to the source-grid-load data grouping results, a generative adversarial network model for each group is constructed and trained, and the source-grid-load data is input into the trained generative adversarial network model to generate a source-grid-load joint scenario set.

[0012] As a preferred solution of the source-grid-load joint scene generation method taking into account meteorological coupling factors described in the present invention, wherein: the correlation analysis of the historical meteorological data and the historical source-grid-load data includes:

[0013] Calculate the correlation value, compare the correlation value with the correlation setting threshold, if the correlation value exceeds the correlation setting threshold, select the corresponding meteorological factor as the key meteorological factor, and calculate the correlation value formula as follows:

[0014]

[0015] Among them, ρ x,y Correlation value, x is the meteorological factor, y is the source-grid-load data, cov(x,y) is the covariance between the meteorological factor and the source-grid-load data, σ x and σ y is the standard deviation of meteorological factors and source-grid-load data, μ x is the average value of meteorological factors, μ y It is the average value of source-grid-load data.

[0016] As a preferred solution of the source-grid-load joint scene generation method taking into account meteorological coupling factors described in the present invention, clustering and grouping the meteorological data according to the key meteorological factors includes:

[0017] Calculating the silhouette coefficient of the key meteorological factor set, and determining the optimal clustering value k based on the silhouette coefficient;

[0018] Select k data points from the key meteorological factor set as the centroid, calculate the Euclidean distance between the remaining key meteorological factors except the centroid and each centroid, and classify the remaining key meteorological factors into the set to which the corresponding centroid belongs according to the nearest distance criterion;

[0019] After all the key meteorological factors are classified into k sets, the centroid of each set is recalculated and replaced. When the distance between the newly calculated centroid and the original centroid is less than the set centroid threshold, the calculation of replacing the centroid is stopped.

[0020] As a preferred solution of the source-grid-load joint scene generation method taking into account meteorological coupling factors described in the present invention, the silhouette coefficient includes:

[0021] The formula for calculating the silhouette coefficient of the key meteorological factor set is:

[0022] SC=(ba) / max{b,a}

[0023] Among them, b is the minimum average distance between a meteorological factor in the key meteorological factor set and a point in other groups, and a is the average distance between a point in the key meteorological factor set and a point in the same group.

[0024] As a preferred solution of the source-grid-load joint scene generation method taking into account meteorological coupling factors described in the present invention, wherein: according to the source-grid-load data grouping results, a generative adversarial network model of each group is constructed and trained, including:

[0025] According to the source-grid-load data grouping result, calculating the true probability distribution of each group of source-grid-load data;

[0026] Generate a predicted probability distribution of source-grid-load data using a generator in a generative adversarial network, and discriminate between a true probability distribution of the source-grid-load data and a predicted probability distribution of the source-grid-load data using a discriminator in the generative adversarial network;

[0027] The generative adversarial network model is continuously adjusted using the minimax game valuation function and the generative adversarial network model loss function.

[0028] As a preferred solution of the source-grid-load joint scenario generation method taking into account meteorological coupling factors described in the present invention, the minimax game valuation function includes:

[0029] Constructing the minimax game valuation function makes the generative adversarial network model reach Nash equilibrium. The minimax game valuation function is expressed as:

[0030] min G max D V(G,D)=E X [D(x)]-E Z [D(G(z))]

[0031] Among them, x is the source-grid-load real data, z is the noise, D(x) and G(z) are the discriminator and generator functions respectively, and E X [·], E Z[·] are the true probability distribution of source-grid-load data and the expectation of random noise distribution, respectively.

[0032] As a preferred solution of the source-grid-load joint scene generation method taking into account meteorological coupling factors described in the present invention, the loss function of the generative adversarial network model includes:

[0033] The Wasserstein distance is used to measure the difference between the true probability distribution of the source-grid-load data and the predicted probability distribution of the source-grid-load data. The loss function of the generative adversarial network model is expressed as:

[0034]

[0035] Where W(D(X),D(G(Z))) is the Wasserstein distance between the true probability distribution of source-grid-load data and the predicted probability distribution of source-grid-load data, sup represents the minimum upper bound, and E X [·], E z [·] are the true probability distribution of source-grid-load data and the expectation of random noise distribution, respectively. D(x) and G(z) are the discriminator and generator functions, respectively.

[0036] In a second aspect, the present invention provides a system for generating a source-grid-load joint scenario taking into account meteorological coupling factors, comprising:

[0037] Data acquisition module, used to obtain historical meteorological data and historical source-grid-load data;

[0038] A cluster analysis module is used to perform correlation analysis on the historical meteorological data and the historical source-grid-load data, obtain key meteorological factors, and cluster the key meteorological factors; according to the cluster grouping results of the key meteorological factors, the source-grid-load data corresponding to the time to which the key meteorological factors belong is divided into groups, and the source-grid-load data grouping results are obtained;

[0039] The scenario generation module is used to construct and train a generative adversarial network model for each group according to the source-grid-load data grouping results, input the source-grid-load data into the trained generative adversarial network model, and generate a source-grid-load joint scenario set.

[0040] In a third aspect, the present invention provides a computing device, comprising:

[0041] Memory and processor;

[0042] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the source-grid-load joint scene generation method taking into account meteorological coupling factors are implemented.

[0043] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the method for generating a source-grid-load joint scenario taking into account meteorological coupling factors.

[0044] Compared with the prior art, the present invention has the following beneficial effects: the present invention jointly generates source-grid-load scenarios based on historical data, without the need for the complex scenario reduction steps required for random matching of scenarios after the "source", "grid" and "load" are generated separately, and the difficulty of establishing a source-grid-load joint probability model is avoided, effectively improving the efficiency of scenario generation. At the same time, meteorological factors are used as the key climate factor set for the correlation between source-grid-load, and a generative adversarial network is used to generate a typical source-grid-load scenario set under different meteorological conditions, which can effectively characterize the volatility and probability distribution characteristics of source, grid and load. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0046] Figure 1 It is a schematic diagram of the overall process of a source-grid-load joint scenario generation method taking into account meteorological coupling factors according to an embodiment of the present invention;

[0047] Figure 2 A schematic diagram of a generative adversarial network model of a source-grid-load joint scenario generation method taking into account meteorological coupling factors according to an embodiment of the present invention;

[0048] Figure 3 A schematic diagram of the Wasserstein distance variation process of a source-grid-load joint scenario generation method taking into account meteorological coupling factors according to an embodiment of the present invention;

[0049] Figure 4 A schematic diagram of a scene generated by a generative adversarial network under different weather types in a source-grid-load joint scene generation method taking into account meteorological coupling factors according to an embodiment of the present invention;

[0050] Figure 5 A schematic diagram of silhouette coefficient values ​​under different clustering numbers of a source-grid-load joint scenario generation method taking into account meteorological coupling factors according to an embodiment of the present invention;

[0051] Figure 6This is a schematic diagram of clustering results when K=4 of a source-grid-load joint scenario generation method taking into account meteorological coupling factors according to an embodiment of the present invention;

[0052] Figure 7 This is a schematic diagram of the distribution of wind power output and load data of each cluster after clustering in a source-grid-load joint scenario generation method taking into account meteorological coupling factors according to an embodiment of the present invention. DETAILED DESCRIPTION

[0053] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0054] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0055] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0056] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0057] At the same time, in the description of the present invention, it should be noted that the orientations or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the orientations or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the system or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0058] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0059] Example 1

[0060] Reference Figure 1-4 , which is an embodiment of the present invention, provides a source-grid-load joint scenario generation method taking into account meteorological coupling factors, comprising:

[0061] S100: Obtain historical meteorological data and historical source-network-load data;

[0062] In the embodiment of the present application, meteorological data includes temperature, humidity, wind speed, wind direction, rainfall, etc., and source-grid-load data covers power supply sources such as power generation of power plants, grid operation conditions such as voltage and current stability, and power loads such as power consumption in different regions or time periods; after obtaining historical meteorological data and historical source-grid-load data, the data is preprocessed, including removing outliers, filling missing values, and standardizing data to ensure data quality and accuracy of analysis. The preprocessed data not only provides a reliable basis for subsequent correlation analysis and clustering grouping, but also directly affects the training effect of the generative adversarial network model and the accuracy and diversity of the final generated source-grid-load joint scenario set.

[0063] S102: Perform correlation analysis on historical meteorological data and historical source-grid-load data to obtain key meteorological factors, and cluster the key meteorological factors;

[0064] Preferably, the correlation value is calculated, and the correlation value is compared with the correlation setting threshold. If the correlation value exceeds the correlation setting threshold, the corresponding meteorological factor is selected as the key meteorological factor. The formula for calculating the correlation value is:

[0065]

[0066] Among them, ρ x,y Correlation value, x is the meteorological factor, y is the source-grid-load data, cov(x,y) is the covariance between the meteorological factor and the source-grid-load data, δ x and σ y is the standard deviation of meteorological factors and source-grid-load data, μ x is the average value of meteorological factors, μ y is the average value of source-grid-load data;

[0067] The larger the correlation value is, the higher the correlation is. Based on this, meteorological factors whose correlation exceeds the correlation setting threshold are extracted, and the key meteorological factor set is constructed using meteorological factors with high correlation with source, network and load.

[0068] It should be noted that setting the correlation threshold is a process that needs to be flexibly adjusted according to actual conditions. The present invention does not specifically limit this. In practical applications, a reasonable threshold can be set based on the statistical analysis results of historical data, expert experience, and specific analysis requirements. This threshold is intended to distinguish between key meteorological factors that have a significant impact on source-grid-load data and factors that have a small or negligible impact, thereby ensuring the effectiveness and accuracy of subsequent analysis. This threshold should be reasonably set according to your own data and goals.

[0069] Preferably, the silhouette coefficient of the key meteorological factor set is calculated, and the optimal clustering value k is determined according to the silhouette coefficient;

[0070] Preferably, k data points are selected from the key meteorological factor set as centroids, the Euclidean distance between the remaining key meteorological factors except the centroid and each centroid is calculated, and the remaining key meteorological factors are classified into the set to which the corresponding centroid belongs according to the closest distance criterion;

[0071] Preferably, after all the key meteorological factors are classified into k sets, the centroid of each set is recalculated and replaced, and when the distance between the newly calculated centroid and the original centroid is less than the set centroid threshold, the calculation of replacing the centroid is stopped;

[0072] When the data tends to converge as a whole, it can be considered that the desired clustering effect has been achieved and the algorithm can be terminated. If the distance between the new centroid and the original centroid changes greatly, it is necessary to repeatedly iterate the centroid calculation until convergence is reached.

[0073] Preferably, the silhouette coefficient formula for calculating the key meteorological factor set is:

[0074] SC=(ba) / max{b,a}

[0075] Among them, b is the minimum average distance between a meteorological factor in the key meteorological factor set and a point in other groups, and a is the average distance between a point in the key meteorological factor set and a point in the same group.

[0076] In the embodiment of the present application, the value range of SC is [-1, +1], and the closer SC is to 1, the better the clustering quality. The present invention uses the average SC of all key meteorological factors as an indicator for evaluating clustering results;

[0077] It should be noted that the present invention effectively extracts key meteorological factors that have a significant impact on the operation of the power system from complex meteorological and source-grid-load data through a systematic data analysis process; first, correlation analysis is used to accurately identify meteorological factors that are closely related to source-grid-load data, and by setting reasonable thresholds, it is ensured that only key factors that have a significant impact on the system are retained; then, the silhouette coefficient is used as an evaluation indicator of the clustering effect, and by continuously optimizing the number of clusters k and the selection of centroids, efficient clustering and grouping of key meteorological factors is achieved. This process not only improves the accuracy of data analysis, but also ensures the stability and reliability of the clustering results, and provides solid data support for subsequent power system optimization scheduling and risk management based on meteorological factors; through this technical solution, the operating efficiency and stability of the power system can be significantly improved, and the potential risks caused by extreme weather can be reduced.

[0078] S104: according to the clustering grouping result of the key meteorological factors, the source-grid-load data corresponding to the time of the key meteorological factors are divided into groups to obtain the source-grid-load data grouping result;

[0079] Specifically, after completing the clustering of key meteorological factors, the analysis results are further applied to the organization and management of historical source-grid-load data. The core is to establish the time correspondence between meteorological factors and source-grid-load data, and divide the source-grid-load data into corresponding meteorological factor groups accordingly;

[0080] First, it is necessary to clarify the timestamp corresponding to each source-grid-load data record. According to the timestamp, each source-grid-load data is matched with the key meteorological factor group that is closest in time. Since the clustering grouping of meteorological factors is based on their historical data patterns, each grouping represents a specific type of meteorological condition or pattern. By matching the source-grid-load data with these meteorological groups, we can better understand the interaction and change patterns between the source-grid-load data under different meteorological conditions.

[0081] During the division process, it is ensured that each source-grid-load data is accurately divided into its corresponding key meteorological factor group, thus forming multiple source-grid-load data grouping results. The data in each group share similar meteorological characteristics, which provides strong support for the subsequent source-grid-load joint scenario modeling based on different meteorological conditions. Through such data division, not only can we have a deeper understanding of the impact of meteorological factors on the operation of the power system, but also provide a more accurate data foundation for the optimal dispatch, risk assessment and decision support of the power system.

[0082] S106: According to the source-network-load data grouping results, a generative adversarial network model for each group is constructed and trained, and the source-network-load data is input into the trained generative adversarial network model to generate a source-network-load joint scenario set; the generative adversarial network model is as follows: Figure 2 As shown;

[0083] Preferably, according to the source-network-load data grouping result, the true probability distribution of each group of source-network-load data is calculated;

[0084] Preferably, a generator in a generative adversarial network is used to generate a predicted probability distribution of source-network-load data, and a discriminator in the generative adversarial network is used to discriminate between the true probability distribution of the source-network-load data and the predicted probability distribution of the source-network-load data;

[0085] Preferably, the generative adversarial network model is continuously adjusted using the minimax game valuation function and the generative adversarial network model loss function;

[0086] In the embodiment of the present application, the generator G includes two deconvolution layers with a stride of 2×2, which upsample the input noise z, and the discriminator D includes two convolution layers with a stride of 2×2, which downsample the scene x; the generator is upsampled from a fully connected multilayer perceptron, and the discriminator has a reverse structure with a single sigmoid output. It was observed experimentally that two convolution layers are sufficient to represent the daily dynamics of the training set, and the training efficiency is high. This embodiment uses the RmsProp optimizer for training, the mini-batch size is 32, and the weights of all structures in the neural network are initialized by Gaussian distribution, with a standard deviation of 0.02. Except for the input layer, batch processing is performed before each layer of the network, the mean and unit variance are set to 0, and the input of each layer is normalized to achieve stable learning; except for the output layer, ReLU activation is used for the generator, Leaky-ReLU activation is used for the discriminator, and ndiscri is set to 4, so that the generative adversarial network is trained alternately between optimizing 4 steps of D and 1 step of g.

[0087] Specifically, in the model training stage, the true probability distribution of each set of source-network-load data is first calculated. In the embodiment of the present application, the true probability distribution of the source-network-load data is obtained by statistical methods, aiming to accurately characterize the inherent laws and characteristics of the data set; the generator part in the generative adversarial network is used to learn and generate the predicted probability distribution of the source-network-load data according to the input data samples. The goal of the generator is to "cheat" the discriminator as much as possible, making it difficult to distinguish the generated data from the real data;

[0088] At the same time, the discriminator in the generative adversarial network plays the role of a verifier, whose responsibility is to distinguish whether the input data comes from the real data set or the fake data generated by the generator. The discriminator continuously optimizes itself to improve its ability to identify the two data sources. In this process, a dynamic game relationship is formed between the generator and the discriminator, the so-called "minimax game", in which the generator tries to minimize the probability of being identified as false by the discriminator, while the discriminator tries to maximize its ability to correctly identify.

[0089] In order to optimize this game process, the Minimax Game Valuation Function is introduced, which takes into account the performance of both the generator and the discriminator and guides the update direction of the network parameters. At the same time, combined with the loss function unique to the generative adversarial network model, the generative adversarial network is continuously adjusted until a balance is reached, that is, the generator can generate sufficiently realistic and indistinguishable source-network-load data samples, and the discriminator's ability to distinguish between real data and generated data reaches a stable level. The trained generative adversarial network model can generate a series of source-network-load joint scene sets.

[0090] Preferably, a minimax game valuation function is constructed to enable the generative adversarial network model to reach a Nash equilibrium. The minimax game valuation function is expressed as:

[0091] min G max D V(G,D)=E X [D(x)]-E Z [D(G(z))]

[0092] Among them, x is the source-grid-load real data, z is the noise, D(x) and G(z) are the discriminator and generator functions respectively, e X [·], E z [·] are the expected probability distribution of source-grid-load data and random noise distribution respectively;

[0093] Preferably, the Wasserstein distance is used to measure the difference between the true probability distribution of source-network-load data and the predicted probability distribution of source-network-load data, and the loss function of the generative adversarial network model is expressed as:

[0094]

[0095] Where W(D(X),D(G(Z))) is the Wasserstein distance between the true probability distribution of source-grid-load data and the predicted probability distribution of source-grid-load data, sup represents the minimum upper bound, and E X [·], E Z[·] are the true probability distribution of source-grid-load data and the expectation of random noise distribution, respectively. D(x) and G(z) are the discriminator and generator functions, respectively.

[0096] The detailed process of the change of the Wasserstein distance of the training data is as follows Figure 3 As shown by Figure 3 It can be seen that at the beginning of training, the samples generated by the generator are far from the real historical data distribution. As the training progresses, the samples generated by the generator gradually approach the real data distribution, and the Wasserstein distance value gradually decreases. After about 20,000 iterations of training, the Wasserstein distance value converges to near 0 and tends to be stable. The training process shows that the samples generated by the generator are very close to the real data distribution and can be used for scene generation.

[0097] 2500 noise vectors Z are extracted from the predefined Gaussian distribution and input into the trained generator. Some generated samples are as follows Figure 4 As shown, it can be seen that the generated scenario has very similar characteristics to the historical data. Therefore, the method of the present invention can capture rich power scenario patterns, thereby generating more accurate scenarios;

[0098] It should be noted that the present invention effectively constructs a model capable of generating a highly realistic source-grid-load joint scenario set through fine source-grid-load data grouping and generative adversarial network model training. In the process of model construction, the true probability distribution of each group of data is accurately calculated by statistical methods, and the minimax game mechanism between the generator and the discriminator of GAN is used to continuously optimize the generator to generate a predicted probability distribution close to the real data distribution. At the same time, the introduction of the Nash equilibrium concept and the Wasserstein distance metric ensures the stability of model training and the accuracy of generated data; the source-grid-load joint scenario set generated by the present invention not only enriches the data foundation of the energy system, but also significantly improves the scientificity and accuracy of energy planning, scheduling and risk management, and provides technical support for the intelligent development of the energy industry.

[0099] The above is a schematic scheme of a source-grid-load joint scene generation method taking into account meteorological coupling factors in this embodiment. It should be noted that the technical scheme of the system for generating a source-grid-load joint scene taking into account meteorological coupling factors and the technical scheme of the above-mentioned source-grid-load joint scene generation method taking into account meteorological coupling factors belong to the same concept. For details not described in detail in the technical scheme of the source-grid-load joint scene generation system taking into account meteorological coupling factors in this embodiment, please refer to the description of the technical scheme of the source-grid-load joint scene generation method taking into account meteorological coupling factors.

[0100] The source-grid-load joint scenario generation system taking into account meteorological coupling factors in this embodiment includes:

[0101] Data acquisition module, used to obtain historical meteorological data and historical source-grid-load data;

[0102] The cluster analysis module is used to perform correlation analysis between historical meteorological data and historical source-grid-load data, obtain key meteorological factors, and cluster the key meteorological factors; according to the clustering grouping results of the key meteorological factors, the source-grid-load data corresponding to the time of the key meteorological factors are divided into groups to obtain the source-grid-load data grouping results;

[0103] The scenario generation module is used to build and train the generative adversarial network model of each group according to the source-grid-load data grouping results, input the source-grid-load data into the trained generative adversarial network model, and generate a source-grid-load joint scenario set.

[0104] This embodiment further provides a computing device, which is applicable to the case of generating a source-grid-load joint scenario taking into account meteorological coupling factors, including:

[0105] A memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the source-grid-load joint scene generation method taking into account meteorological coupling factors as proposed in the above embodiment.

[0106] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for generating a source-grid-load joint scenario taking into account meteorological coupling factors as proposed in the above embodiment is implemented.

[0107] The storage medium proposed in this embodiment and the method for generating a source-grid-load joint scene taking into account meteorological coupling factors proposed in the above embodiment belong to the same inventive concept. For technical details not described in detail in this embodiment, please refer to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0108] Through the above description of the implementation methods, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ReadOnly, Memory, ROM), random access memory (RandomAccess Memory, RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform the methods of various embodiments of the present invention.

[0109] Example 2

[0110] Reference Figure 5-7 Table 1 is an embodiment of the present invention, which provides a source-grid-load joint scenario generation method taking into account meteorological coupling factors. In order to verify its beneficial effects, the experimental verification results of the present invention are provided.

[0111] This embodiment uses the measured historical data of a wind farm for 365 consecutive days from January 1 to December 31, 2017, including temperature, humidity, wind speed, wind power output, load, etc. to perform meteorological coupling clustering and source-load joint scenario generation. The following studies are all based on this data set.

[0112] The correlation between meteorological factors and source-grid-load data is calculated, as shown in Table 1. Table 1 shows the correlation between meteorological factors such as temperature, humidity, wind speed, and wind power output and load;

[0113] Table 1 Correlation between wind power output, load and meteorological factors

[0114]

[0115] As shown in Table 1, different meteorological factors have certain correlations with wind power output and load, but the correlations of various factors are different; among them, wind speed, humidity and wind power output, as well as temperature, humidity and load have high correlations. Although the correlation between air pressure and wind power output and load is low, it is still a key meteorological factor.

[0116] Therefore, meteorological factors such as temperature, humidity, and wind speed are used to construct a set of key meteorological factors.

[0117] The average silhouette coefficient value of all elements in the key meteorological factor set is used as the evaluation index of the clustering result. By comparing the silhouette coefficient values ​​of different numbers of clusters, it is found that the clustering effect is most suitable when K = 4, such as Figure 5As shown in the figure, this clustering number selection method optimizes the silhouette coefficient, which can effectively characterize the similarity within a class and the difference between classes, and is particularly important for accurate scene generation.

[0118] Figure 6 This is the best clustering result of annual meteorological data. The four colors represent the clusters after clustering.

[0119] After clustering based on meteorological coupling feature sets, the wind power output and load data distribution of each cluster and the cluster center are as follows: Figure 7 As shown by Figure 7 It can be seen that the data within each class are mainly concentrated near the center, with few outliers;

[0120] There are significant differences in peak values, volatility, and change trends in different classes. For example, there is an obvious gap in wind power output in group 3, which shows the intermittent nature of wind power output under certain meteorological conditions. Therefore, the clustering method based on key meteorological factors used in the present invention can effectively characterize the intra-class similarities and inter-class differences of wind power output and load.

[0121] By deeply analyzing the correlation between meteorological factors and wind power output and load, and conducting cluster analysis based on a set of key meteorological factors, the wind power output and load characteristics under different meteorological conditions are effectively identified and characterized, providing more accurate data support for the generation of source-load joint scenarios, and significantly improving the accuracy and practicality of scenario generation.

[0122] 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 the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A source-grid-load joint scenario generation method taking into account meteorological coupling factors, characterized in that: include: Obtain historical meteorological data and historical source-grid-load data; Perform correlation analysis on the historical meteorological data and historical source-grid-load data to obtain key meteorological factors, and cluster and group the key meteorological factors; According to the clustering grouping results of key meteorological factors, the source-grid-load data corresponding to the time of the key meteorological factors are divided into groups to obtain the source-grid-load data grouping results; According to the source-grid-load data grouping results, a generative adversarial network model for each group is constructed and trained, and the source-grid-load data is input into the trained generative adversarial network model to generate a source-grid-load joint scenario set.

2. The source-grid-load joint scenario generation method taking into account meteorological coupling factors according to claim 1, characterized in that: The correlation analysis of the historical meteorological data and the historical source-grid-load data includes: Calculate the correlation value, compare the correlation value with the correlation setting threshold, if the correlation value exceeds the correlation setting threshold, select the corresponding meteorological factor as the key meteorological factor, the formula for calculating the correlation value is: Among them, v x,y Correlation value, x is the meteorological factor, y is the source-grid-load data, cov(x,y) is the covariance between the meteorological factor and the source-grid-load data, σ x and σ y is the standard deviation of meteorological factors and source-grid-load data, μ x is the average value of meteorological factors, μ y It is the average value of source-grid-load data.

3. The source-grid-load joint scenario generation method taking into account meteorological coupling factors according to claim 1 or 2, characterized in that: Clustering the meteorological data according to the key meteorological factors includes: Calculating the silhouette coefficient of the key meteorological factor set, and determining the optimal clustering value k based on the silhouette coefficient; Select k data points from the key meteorological factor set as the centroid, calculate the Euclidean distance between the remaining key meteorological factors except the centroid and each centroid, and classify the remaining key meteorological factors into the set to which the corresponding centroid belongs according to the nearest distance criterion; After all the key meteorological factors are classified into k sets, the centroid of each set is recalculated and replaced. When the distance between the newly calculated centroid and the original centroid is less than the set centroid threshold, the calculation of replacing the centroid is stopped.

4. The source-grid-load joint scenario generation method taking into account meteorological coupling factors as claimed in claim 3, characterized in that: The silhouette coefficient includes, The formula for calculating the silhouette coefficient of the key meteorological factor set is: SC=(ba) / max{b,a} Among them, b is the minimum average distance between a meteorological factor in the key meteorological factor set and a point in other groups, and a is the average distance between a point in the key meteorological factor set and a point in the same group.

5. The source-grid-load joint scenario generation method taking into account meteorological coupling factors as claimed in claim 4, characterized in that: According to the source-network-load data grouping results, a generative adversarial network model for each group is constructed and trained, including: According to the source-grid-load data grouping result, calculating the true probability distribution of each group of source-grid-load data; Generate a predicted probability distribution of source-grid-load data using a generator in a generative adversarial network, and discriminate between a true probability distribution of the source-grid-load data and a predicted probability distribution of the source-grid-load data using a discriminator in the generative adversarial network; The generative adversarial network model is continuously adjusted using the minimax game valuation function and the generative adversarial network model loss function.

6. The source-grid-load joint scenario generation method taking into account meteorological coupling factors as claimed in claim 5, characterized in that: Minimax game valuation functions include: Constructing the minimax game valuation function makes the generative adversarial network model reach Nash equilibrium. The minimax game valuation function is expressed as: min G max D V(G,D)=E X [D(x)]-E Z [D(G(z))] Among them, x is the source-grid-load real data, z is the noise, D(x) and G(z) are the discriminator and generator functions respectively, and E X [·], E Z [·] are the true probability distribution of source-grid-load data and the expectation of random noise distribution, respectively.

7. The source-grid-load joint scenario generation method taking into account meteorological coupling factors as claimed in claim 5, characterized in that: The loss function of the generative adversarial network model includes: The Wasserstein distance is used to measure the difference between the true probability distribution of the source-grid-load data and the predicted probability distribution of the source-grid-load data. The loss function of the generative adversarial network model is expressed as: Where W(D(X),D(G(Z))) is the Wasserstein distance between the true probability distribution of source-grid-load data and the predicted probability distribution of source-grid-load data, sup represents the minimum upper bound, and E X [·], E Z [·] are the true probability distribution of source-grid-load data and the expectation of random noise distribution, d(x) and G(z) are the discriminator and generator functions, respectively.

8. A system for generating source-grid-load joint scenarios taking into account meteorological coupling factors, characterized in that: include, Data acquisition module, used to obtain historical meteorological data and historical source-grid-load data; A cluster analysis module is used to perform correlation analysis on the historical meteorological data and the historical source-grid-load data, obtain key meteorological factors, and cluster the key meteorological factors; according to the cluster grouping results of the key meteorological factors, the source-grid-load data corresponding to the time to which the key meteorological factors belong is divided into groups, and the source-grid-load data grouping results are obtained; The scenario generation module is used to construct and train a generative adversarial network model for each group according to the source-grid-load data grouping results, input the source-grid-load data into the trained generative adversarial network model, and generate a source-grid-load joint scenario set.

9. An electronic device, comprising: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the source-grid-load joint scene generation method taking into account meteorological coupling factors as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the source-grid-load joint scenario generation method taking into account meteorological coupling factors as described in any one of claims 1 to 7.

Images