Methods, devices, and non-volatile storage media for dispatching discrete-action equipment in power distribution networks

Abstract

This application discloses a method, apparatus, and non-volatile storage medium for scheduling discrete-operated equipment in a distribution network. The method includes: determining load forecast data and active power output forecast data of the distributed generation in the distribution network within a preset time period; processing the load forecast data and active power output forecast data using a discrete-operated equipment scheduling model to obtain a scheduling scheme for the discrete-operated equipment in the distribution network. The training data for the discrete-operated equipment scheduling model includes labeled and unlabeled data, and the model includes a feature extractor for identifying the data input to the model; adjusting the tap switching positions of the discrete-operated equipment within the preset time period according to the scheduling scheme. This application solves the technical problem of the inability to automatically schedule discrete-operated equipment in a distribution network due to inaccurate output results from discrete-operated equipment scheduling models in related technologies.

Description

Technical Field

[0001] This application relates to the field of electrical engineering, and more specifically, to a method, apparatus, and non-volatile storage medium for scheduling discrete-action equipment in a power distribution network. Background Technology

[0002] To ensure the accuracy of the scheduling results, discrete-action equipment scheduling models in related technologies rely on a large amount of reliable labeled data for training, rather than using unlabeled data. However, data in distribution networks often contains a large amount of unlabeled data, making it impossible to guarantee the accuracy of the output results of discrete-action equipment scheduling models in related technologies, thus hindering the automated scheduling of discrete-action equipment in distribution networks.

[0003] There is currently no effective solution to the above problems. Summary of the Invention

[0004] This application provides a method, apparatus, and non-volatile storage medium for scheduling discrete-action equipment in a distribution network, to at least solve the technical problem that the discrete-action equipment in the distribution network cannot be automatically scheduled due to inaccurate output results of the discrete-action equipment scheduling model in related technologies.

[0005] According to one aspect of the embodiments of this application, a method for scheduling discrete-operated equipment in a distribution network is provided, comprising: determining load forecast data of the distribution network and active power output forecast data of distributed generation in the distribution network within a preset time period; calling a discrete-operated equipment scheduling model to process the load forecast data and active power output forecast data, thereby obtaining a scheduling scheme for discrete-operated equipment in the distribution network, wherein the scheduling scheme includes the predicted switching positions of the taps of the discrete-operated equipment within the preset time period, the training data of the discrete-operated equipment scheduling model includes labeled data and unlabeled data, the discrete-operated equipment scheduling model includes a feature extractor for identifying the data input into the discrete-operated equipment scheduling model, and an estimation module for generating the scheduling scheme; and adjusting the tap switching positions of the discrete-operated equipment within the preset time period according to the scheduling scheme.

[0006] Optionally, the discrete motion equipment scheduling model is invoked to process load forecast data and active power output forecast data to obtain a scheduling scheme for discrete motion equipment in the distribution network. This includes: extracting data features from the load forecast data and active power output forecast data through the feature extractor in the discrete motion equipment scheduling model, and performing data dimensionality reduction and global average pooling on the data features to obtain a data feature vector of a preset length; and determining the location prediction value of the data feature vector mapping through the estimation module in the equipment scheduling model via a fully connected layer, wherein the location prediction value is used to indicate the predicted switching position.

[0007] Optionally, the discrete-action equipment scheduling model is trained in the following manner: acquiring historical data of the distribution network, including historical load forecast data and historical active power output forecast data; identifying labeled and unlabeled historical data; determining the principal component features of the labeled historical data and filtering the labeled historical data based on the principal component features to obtain a first training dataset; using the principal component features of the labeled historical data as reference principal component features for the unlabeled historical data and performing clustering filtering on the unlabeled historical data based on the reference principal component features to obtain a second training dataset; and using a generative adversarial network to train the discrete-action equipment scheduling model based on the first and second training datasets.

[0008] Optionally, determining the principal component features of the labeled historical data includes: standardizing the labeled data so that the mean of the features is a first preset value and the variance is a second preset value; determining the covariance matrix of the standardized labeled data, wherein the principal component features include principal component eigenvectors; the covariance matrix is ​​used to reflect the interrelationships between the various features in the labeled data; performing eigenvalue decomposition on the covariance matrix to obtain eigenvalues ​​and their corresponding eigenvectors; and determining the principal component eigenvectors in the labeled data based on the magnitude of the eigenvalues.

[0009] Optionally, the process of filtering labeled historical data based on principal component features to obtain the first training dataset includes: determining the correlation between principal component features and node voltage features of the distribution network; and filtering labeled data based on the correlation to obtain the first training dataset.

[0010] Optionally, the second training dataset is obtained by clustering and filtering unlabeled historical data based on the reference principal component features, including: standardizing the unlabeled data so that the feature mean of the unlabeled data is a first preset value and the variance is a second preset value; constructing a feature space based on the interrelationships reflected in the covariance matrix and the standardized unlabeled data; determining the principal component space based on the reference principal components, and processing the feature space in the principal component space to obtain the feature representation corresponding to the unlabeled historical data; performing cluster analysis on the feature representation to obtain the second training dataset, and verifying the influence mode and degree of the tap position of discrete action equipment in the unlabeled data on the voltage control result during the cluster analysis process.

[0011] Optionally, the method further includes: during the training process, determining the cross-entropy loss of the discrete action equipment scheduling model, and determining the loss function value of the discrete action equipment scheduling model and the gear switching penalty coefficient of the discrete action equipment in the distribution network based on the cross-entropy loss, wherein the gear switching penalty coefficient is used to adjust the loss function value, and the larger the gear switching penalty coefficient is, the larger the adjusted loss function value is.

[0012] According to another aspect of the embodiments of this application, a distribution network discrete-operation equipment scheduling device is also provided, comprising: a first processing module, configured to determine load forecast data of the distribution network and active power output forecast data of the distributed generation of the distribution network within a preset time period; a second processing module, configured to call a discrete-operation equipment scheduling model to process the load forecast data and active power output forecast data, thereby obtaining a scheduling scheme for discrete-operation equipment in the distribution network, wherein the scheduling scheme includes the predicted switching positions of the taps of the discrete-operation equipment within the preset time period, the training data of the discrete-operation equipment scheduling model includes labeled data and unlabeled data, the discrete-operation equipment scheduling model includes a feature extractor for identifying the data input into the discrete-operation equipment scheduling model, and an estimation module for generating the scheduling scheme; and a third processing module, configured to adjust the tap switching positions of the discrete-operation equipment within the preset time period according to the scheduling scheme.

[0013] According to another aspect of the embodiments of this application, a non-volatile storage medium is also provided, wherein a program is stored in the non-volatile storage medium, and the program controls the device where the non-volatile storage medium is located to execute a power distribution network discrete action device scheduling method when it runs.

[0014] According to another aspect of the embodiments of this application, an electronic device is also provided, including: a memory and a processor, wherein the processor is used to run a program stored in the memory, wherein the program executes a method for scheduling discrete action devices in a power distribution network during runtime.

[0015] According to another aspect of the embodiments of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements a method for scheduling discrete action devices in a power distribution network.

[0016] In this embodiment, load forecast data of the distribution network and active power output forecast data of the distributed generation of the distribution network are determined within a preset time period. A discrete motion equipment scheduling model is invoked to process the load forecast data and active power output forecast data, thereby obtaining a scheduling scheme for discrete motion equipment in the distribution network. The scheduling scheme includes the predicted switching positions of the taps of the discrete motion equipment within the preset time period. The training data of the discrete motion equipment scheduling model includes labeled and unlabeled data. The discrete motion equipment scheduling model includes a feature extractor for identifying the data input to the model and an estimation module for generating the scheduling scheme. The switching positions of the taps of the discrete motion equipment within the preset time period are adjusted according to the scheduling scheme. By using both labeled and unlabeled data to train the discrete motion equipment scheduling model, the accuracy of the model output results can still be guaranteed even when the input data is low-quality unlabeled data. This achieves the technical effect of automated scheduling of discrete motion equipment, thus solving the technical problem that the inaccurate output results of discrete motion equipment scheduling models in related technologies prevent automated scheduling of discrete motion equipment in the distribution network. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0018] Figure 1 This is a schematic diagram of the structure of a computer terminal (mobile terminal) according to an embodiment of this application;

[0019] Figure 2 This is a flowchart illustrating a method for scheduling discrete action devices in a power distribution network according to an embodiment of this application.

[0020] Figure 3 This is a schematic diagram illustrating the training process of a semi-supervised distribution network discrete action equipment scheduling model provided in an embodiment of this application;

[0021] Figure 4 This is a schematic diagram of the structure of a power distribution network discrete action equipment scheduling device according to an embodiment of this application. Detailed Implementation

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

[0023] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0024] With the gradual advancement of the "carbon peaking and carbon neutrality" goals, the penetration rate of distributed power sources, mainly wind and solar, in distribution networks is continuously increasing, posing a severe challenge to the safe and economical operation of these networks. Traditional day-ahead dispatching of discrete-operation equipment requires a large amount of high-quality labeled data. However, many distribution networks suffer from missing or inaccurate historical data records, leading to unreasonable day-ahead dispatching results from the models. Traditional day-ahead dispatching of discrete-operation equipment in distribution networks typically relies heavily on a large amount of reliable labeled data. Due to the difficulty in acquiring and maintaining data and the uncertainty of data quality, the reliability of traditional day-ahead dispatching models for discrete-operation equipment in distribution networks is relatively weak. Furthermore, because different networks have their unique load characteristics and operating features, traditional day-ahead dispatching models for discrete-operation equipment in distribution networks have fixed patterns and poor flexibility.

[0025] To address these issues, this application provides a day-ahead scheduling model and training method for discrete-action equipment in distribution networks based on a semi-supervised adversarial deep learning approach. The semi-supervised adversarial deep learning model has low requirements for labeled data, effectively solving the data requirements and labeling problems of traditional day-ahead scheduling models for discrete-action equipment in distribution networks. Furthermore, the proposed day-ahead scheduling model for discrete-action equipment in distribution networks based on the semi-supervised adversarial deep learning approach also incorporates a historical data evaluation module. This module can filter out high-quality data from a large amount of historical data of varying quality for model training, solving the problem of poor model performance due to high dependence on data quality. This will be explained in detail below.

[0026] According to an embodiment of this application, a method embodiment for scheduling discrete action devices in a power distribution network is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0027] The method embodiments provided in this application can be executed on a mobile terminal, computer terminal, or similar computing device. Figure 1 A hardware block diagram of a computer terminal (or mobile device) for implementing a method for scheduling discrete-action devices in a power distribution network is shown. Figure 1 As shown, the computer terminal 10 (or mobile device 10) may include one or more processors 102 (shown as 102a, 102b, ..., 102n in the figure) 102 (processor 102 may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, computer terminal 10 may also include... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0028] It should be noted that the aforementioned one or more processors 102 and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element within the computer terminal 10 (or mobile device). As involved in the embodiments of this application, the data processing circuits serve as a processor control mechanism (e.g., selection of a variable resistor termination path connected to an interface).

[0029] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the power distribution network discrete action equipment scheduling method in this embodiment. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby realizing the aforementioned power distribution network discrete action equipment scheduling method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the computer terminal 10 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0030] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the computer terminal 10. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.

[0031] The display can be, for example, a touchscreen liquid crystal display (LCD) that allows the user to interact with the user interface of the computer terminal 10 (or mobile device).

[0032] Under the above operating environment, embodiments of this application provide a method for scheduling discrete-action devices in a power distribution network, such as... Figure 2 As shown, the method includes the following steps:

[0033] Step S202: Determine the load forecast data of the distribution network and the active power output forecast data of the distributed generation of the distribution network within a preset time period.

[0034] It should be noted that day-ahead dispatching of discrete-action equipment is an important technique implemented in distribution networks, aiming to improve voltage quality and reduce active power losses. Its objective is to minimize active power losses and voltage deviations. The minimum expected active power loss f1 can be expressed as follows:

[0035]

[0036] In the formula, N is the total number of nodes in the distribution network; R ij Let n be the resistance of branch ij; s Number of scenes; Iij,t Let represent the magnitude of the current flowing through branch ij during time period t in scenario s.

[0037] The desired minimum voltage offset f2 can be expressed as follows:

[0038]

[0039] In the formula U i,s,t U represents the voltage amplitude at node i during time period t in scenario s; i,max U i,min U r The upper and lower voltage limits and the rated voltage of node i are, in order.

[0040] In some embodiments of this application, the following constraints should also be satisfied during the day-ahead scheduling of discrete-action equipment: active power and reactive power balance constraints; voltage and current constraints; substation outlet power constraints; control variable constraints; and constraints on the number of operations of capacitor banks and on-load tap-changing transformers.

[0041] Step S204: Call the discrete motion equipment scheduling model to process load forecast data and active power output forecast data, thereby obtaining the scheduling scheme of discrete motion equipment in the distribution network. The scheduling scheme includes the predicted switching positions of the taps of discrete motion equipment within a preset time period. The training data of the discrete motion equipment scheduling model includes labeled data and unlabeled data. The discrete motion equipment scheduling model includes a feature extractor for identifying the data input into the discrete motion equipment scheduling model, and an estimation module for generating the scheduling scheme.

[0042] It should be noted that the above-described discrete-action equipment scheduling model is a day-ahead scheduling model for discrete-action equipment. This model mainly includes two modules: a historical data quality assessment and screening module, and a semi-supervised adversarial deep learning module for day-ahead scheduling. The historical data quality assessment and screening module is used to select high-quality data from a large amount of historical data for model training. The trained semi-supervised adversarial deep learning module performs day-ahead scheduling of discrete-action equipment (on-load tap-changing transformers, parallel capacitor banks) in the distribution network based on load and distributed generation forecast data. The trained semi-supervised adversarial deep learning module includes a feature extractor and an estimation module.

[0043] In the technical solution provided in step S204, the discrete motion equipment scheduling model is called to process the load forecast data and active power output forecast data to obtain the scheduling scheme of discrete motion equipment in the distribution network. This includes: extracting data features from the load forecast data and active power output forecast data through the feature extractor in the discrete motion equipment scheduling model, and performing data dimensionality reduction and global average pooling on the data features to obtain a data feature vector of a preset length; and determining the location prediction value of the data feature vector mapping through the estimation module in the equipment scheduling model via the fully connected layer, wherein the location prediction value is used to indicate the predicted switching position.

[0044] As an optional implementation method, such as Figure 3 As shown, the discrete motion equipment scheduling model is trained in the following way: acquiring historical data of the distribution network, including historical load forecast data and historical active power output forecast data; identifying labeled and unlabeled historical data; determining the principal component features of the labeled historical data and filtering the labeled historical data based on the principal component features to obtain the first training dataset; using the principal component features of the labeled historical data as reference principal component features for the unlabeled historical data and clustering the unlabeled historical data based on the reference principal component features to obtain the second training dataset; and using a generative adversarial network to train the discrete motion equipment scheduling model based on the first and second training datasets.

[0045] In some embodiments of this application, the step of determining the principal component features of labeled historical data includes: standardizing the labeled data so that the feature mean of the labeled data is a first preset value and the variance is a second preset value; determining the covariance matrix of the standardized labeled data, wherein the principal component features include principal component eigenvectors; the covariance matrix is ​​used to reflect the interrelationships between the various feature quantities in the labeled data; performing eigenvalue decomposition on the covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors; and determining the principal component eigenvectors in the labeled data based on the magnitude of the eigenvalues.

[0046] As an optional implementation method, the step of filtering labeled historical data based on principal component features to obtain the first training dataset includes: determining the correlation between principal component features and node voltage features of the distribution network; and filtering labeled data based on the correlation to obtain the first training dataset.

[0047] Specifically, such as Figure 3 As shown, the specific process for filtering labeled historical data is as follows:

[0048] The first step is standardization: Standardize the labeled dataset so that the mean of each feature is 0 and the variance is 1.

[0049] The second step is to calculate the covariance matrix: the covariance matrix is ​​calculated using the standardized data to obtain the relationships between the various features.

[0050] The third step is eigenvalue decomposition: Eigenvalue decomposition is performed on the covariance matrix to extract eigenvalues ​​and eigenvectors.

[0051] The fourth step is to select principal components: based on the magnitude of the eigenvalues, select the principal components.

[0052] The fifth step is correlation analysis: analyzing the correlation between principal components and node voltage features, and selecting high-quality historical data as labeled datasets for training.

[0053] As an optional implementation, when determining the correlation between principal component features and the node voltage characteristics of the distribution network, the Pearson correlation coefficient between the principal component features and the node voltage characteristics can be calculated to quantify the degree of association between each principal component feature and the node voltage characteristics. A Pearson correlation coefficient threshold can then be set, and labeled data exceeding this threshold can be retained in the first training dataset. Higher correlation data is generally considered to have higher data quality.

[0054] In some embodiments of this application, the step of clustering and filtering unlabeled historical data based on reference principal component features to obtain a second training dataset includes: standardizing the unlabeled data so that the feature mean of the unlabeled data is a first preset value and the variance is a second preset value; constructing a feature space based on the interrelationships reflected in the covariance matrix and the standardized unlabeled data; determining the principal component space based on the reference principal components, and processing the feature space in the principal component space to obtain a feature representation corresponding to the unlabeled historical data; performing cluster analysis on the feature representation to obtain a second training dataset, and verifying the influence mode and degree of the tap position of discrete motion equipment in the unlabeled data on the voltage control result during the cluster analysis process.

[0055] Specifically, such as Figure 3 As shown, the filtering process for unlabeled data is as follows:

[0056] The first step is standardization: Standardize the unlabeled dataset so that the mean of each feature is 0 and the variance is 1.

[0057] The second step is to determine the reference principal components: obtain the principal component analysis results from the labeled dataset.

[0058] The third step is feature relationship mapping: the feature relationships obtained from the labeled dataset are mapped to the unlabeled dataset to form a new feature space.

[0059] The fourth step is to project the data into the principal component space: analyze the data in the unlabeled dataset in the principal component space formed by the supervised dataset to obtain new feature representations.

[0060] The fifth step is cluster analysis: cluster analysis is performed on the new feature representation to verify the effectiveness of the tap position of discrete motion devices in voltage control in the unlabeled dataset.

[0061] Specifically, the covariance matrix can be calculated using standardized unlabeled data. Then, eigenvalue decomposition is performed on the covariance matrix to obtain eigenvalues ​​and their corresponding eigenvectors. The eigenvectors are then sorted and selected according to the magnitude of their eigenvalues, with the eigenvectors having a cumulative variance of 90% selected as the principal features. Finally, the unlabeled data can be projected onto the selected principal component directions to obtain the dimensionality-reduced feature representation.

[0062] In some embodiments of this application, the semi-supervised adversarial deep learning module is used to perform day-ahead scheduling of discrete-action equipment based on the day-ahead 24-hour load forecast data and the day-ahead 24-hour distributed power generation active power output forecast data. The semi-supervised adversarial deep learning module mainly consists of two parts: semi-supervised deep learning and an adversarial generative network (GAN). The semi-supervised deep learning is the main part of the model, including a feature extractor for identifying the input day-ahead forecast data and an estimation module for estimating the tap changer positions of discrete-action equipment; the GAN is used to improve the feature extraction capability of the feature extractor and the generalization capability of the model. The specific operation flow of the semi-supervised deep learning module is as follows: Figure 3 As shown, it includes the following steps:

[0063] The first step is to load and preprocess the data. From the reasonable dataset selected by the quality assessment and screening module of historical data, the data is divided into labeled datasets and unlabeled datasets. Data processing is then performed to map the tap positions of discrete motion devices to a positive integer range and convert these integer labels into one-hot encoding format to adapt to the neural network classification task.

[0064] The second step involves using a feature extractor to extract data features. After extracting temporal features from the input data using a three-layer one-dimensional convolutional layer, the data dimensionality is reduced. Finally, a global average pooling layer maps the features into a fixed-length vector.

[0065] The third step involves pre-training the model using labeled data. In each iteration, the feature extractor transforms the input data into feature vectors, which are then mapped by the estimation module through a fully connected layer to the estimated shift position of the discrete motion device. Finally, the loss and the penalty for gear shifting are calculated based on the cross-entropy loss.

[0066] The operational flow of the Generative Adversarial Network (GAN) is as follows: Figure 3 As shown, it includes the following steps:

[0067] The first step is to train the discriminator, which is composed of a fully connected neural network. The discriminator is input with the data features extracted by the feature extraction module, and the discriminator outputs a probability indicating whether the data comes from real data or the generated network.

[0068] The second step involves adversarial training of the feature extractor and discriminator using both unlabeled and labeled data. The feature extractor is responsible for extracting features from both the unlabeled and labeled data, while the discriminator determines whether the extracted features originate from unlabeled or labeled data. During training, the discriminator loss is calculated and backpropagation is used to update the discriminator parameters, ultimately improving the feature extraction capability of the feature extractor.

[0069] As an optional implementation, during the training process, the cross-entropy loss of the discrete motion equipment scheduling model can also be determined, and the loss function value of the discrete motion equipment scheduling model and the gear switching penalty coefficient of the discrete motion equipment in the distribution network can be determined based on the cross-entropy loss. The gear switching penalty coefficient is used to adjust the loss function value, and the larger the gear switching penalty coefficient is, the larger the adjusted loss function value is.

[0070] Specifically, during the training process, the action of the discrete action device corresponding to the output result can be determined based on the model's output result. Then, when the action result meets the constraint conditions or is far from the control target, a larger penalty coefficient is given to the model to increase the loss function value of the model.

[0071] Step S206: Adjust the tap switching position of the discrete action equipment within a preset time period according to the scheduling scheme.

[0072] By employing load forecast data and active power output forecast data of the distributed generation in the distribution network within a predetermined time period, and by calling a discrete-action equipment scheduling model to process the load forecast data and active power output forecast data, a scheduling scheme for discrete-action equipment in the distribution network is obtained. The scheduling scheme includes the predicted switching positions of the taps of the discrete-action equipment within the predetermined time period. The training data for the discrete-action equipment scheduling model includes labeled and unlabeled data. The discrete-action equipment scheduling model includes a feature extractor for identifying the data input to the model and an estimation module for generating the scheduling scheme. The switching positions of the taps of the discrete-action equipment within the predetermined time period are adjusted according to the scheduling scheme. By using both labeled and unlabeled data to train the discrete-action equipment scheduling model, the accuracy of the model output results can still be guaranteed even when the input data is low-quality unlabeled data. This achieves the technical effect of automated scheduling of discrete-action equipment, thus solving the technical problem of the inability to automatically schedule discrete-action equipment in the distribution network due to inaccurate output results of discrete-action equipment scheduling models in related technologies.

[0073] Furthermore, the discrete action model scheduling method for distribution networks provided in this application utilizes a semi-supervised deep learning algorithm, reducing the need for labeled data. It employs a large amount of unlabeled data supplemented by a small amount of labeled data for training, making it more suitable for day-ahead scheduling of large-scale distribution networks with massive amounts of data. Moreover, the model training process incorporates a generative adversarial network, resulting in stronger feature extraction capabilities and making it more suitable for distribution networks with complex and variable operating conditions.

[0074] This application provides a dispatching device for discrete action equipment in a power distribution network. Figure 4 This is a schematic diagram of the device. From Figure 4 As can be seen from the diagram, the device includes: a first processing module 40, used to determine the load forecast data of the distribution network and the active power output forecast data of the distributed power sources in the distribution network within a preset time period; a second processing module 42, used to call the discrete motion equipment scheduling model to process the load forecast data and active power output forecast data, thereby obtaining the scheduling scheme of discrete motion equipment in the distribution network, wherein the scheduling scheme includes the predicted switching positions of the taps of the discrete motion equipment within the preset time period, the training data of the discrete motion equipment scheduling model includes labeled data and unlabeled data, the discrete motion equipment scheduling model includes a feature extractor for identifying the data input into the discrete motion equipment scheduling model, and an estimation module for generating the scheduling scheme; and a third processing module 44, used to adjust the tap switching positions of the discrete motion equipment within the preset time period according to the scheduling scheme.

[0075] In some embodiments of this application, the second processing module 42 calls the discrete motion equipment scheduling model to process load forecast data and active power output forecast data, thereby obtaining the scheduling scheme of discrete motion equipment in the distribution network. The steps include: extracting data features from the load forecast data and active power output forecast data through the feature extractor in the discrete motion equipment scheduling model, and performing data dimensionality reduction and global average pooling on the data features to obtain a data feature vector of a preset length; and determining the location prediction value of the data feature vector mapping through the fully connected layer by the estimation module in the equipment scheduling model, wherein the location prediction value is used to indicate the predicted switching position.

[0076] In some embodiments of this application, the discrete motion equipment scheduling model is trained in the following manner: acquiring historical data of the distribution network, wherein the historical data includes historical load forecast data and historical active power output forecast data; determining labeled historical data and unlabeled historical data in the historical data; determining the principal component features of the labeled historical data, and filtering the labeled historical data according to the principal component features to obtain a first training dataset; using the principal component features of the labeled historical data as reference principal component features of the unlabeled historical data, and performing clustering filtering on the unlabeled historical data according to the reference principal component features to obtain a second training dataset; and using a generative adversarial network to train the discrete motion equipment scheduling model based on the first training dataset and the second training dataset.

[0077] In some embodiments of this application, determining the principal component features of labeled historical data includes: standardizing the labeled data so that the mean of the features of the labeled data is a first preset value and the variance is a second preset value; determining the covariance matrix of the standardized labeled data, wherein the principal component features include principal component eigenvectors; the covariance matrix is ​​used to reflect the interrelationships between the various features in the labeled data; performing eigenvalue decomposition on the covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors; and determining the principal component eigenvectors in the labeled data based on the magnitude of the eigenvalues.

[0078] In some embodiments of this application, the step of filtering labeled historical data based on principal component features to obtain a first training dataset includes: determining the correlation between principal component features and node voltage features of the distribution network; filtering labeled data based on the correlation to obtain a first training dataset.

[0079] In some embodiments of this application, the step of clustering and filtering unlabeled historical data based on reference principal component features to obtain a second training dataset includes: standardizing the unlabeled data so that the feature mean of the unlabeled data is a first preset value and the variance is a second preset value; constructing a feature space based on the interrelationships reflected in the covariance matrix and the standardized unlabeled data; determining the principal component space based on the reference principal components, and processing the feature space in the principal component space to obtain a feature representation corresponding to the unlabeled historical data; performing cluster analysis on the feature representation to obtain a second training dataset, and verifying the influence mode and degree of the tap position of discrete motion equipment in the unlabeled data on the voltage control result during the cluster analysis process.

[0080] In some embodiments of this application, the discrete action equipment scheduling device for the distribution network is further configured to: determine the cross-entropy loss of the discrete action equipment scheduling model during the training process, and determine the loss function value of the discrete action equipment scheduling model based on the cross-entropy loss, as well as the gear switching penalty coefficient of the discrete action equipment in the distribution network, wherein the gear switching penalty coefficient is used to adjust the loss function value, and the larger the gear switching penalty coefficient is, the larger the adjusted loss function value is.

[0081] It should be noted that each module in the aforementioned power distribution network discrete action equipment scheduling device can be a program module (e.g., a set of program instructions to implement a specific function) or a hardware module. For the latter, it can be expressed in the following forms, but is not limited to them: each of the above modules is expressed as a processor, or the functions of each of the above modules are implemented by a processor.

[0082] According to an embodiment of this application, a non-volatile storage medium is also provided, which stores a program. During program execution, the device containing the non-volatile storage medium executes the following method for scheduling discrete-action devices in a distribution network: determining load forecast data and active power output forecast data of distributed generation sources within a preset time period; calling a discrete-action device scheduling model to process the load forecast data and active power output forecast data to obtain a scheduling scheme for discrete-action devices in the distribution network. The scheduling scheme includes the predicted switching positions of the taps of the discrete-action devices within the preset time period. The training data for the discrete-action device scheduling model includes labeled and unlabeled data. The discrete-action device scheduling model includes a feature extractor for identifying the data input to the model and an estimation module for generating the scheduling scheme; and adjusting the tap switching positions of the discrete-action devices within the preset time period according to the scheduling scheme.

[0083] According to an embodiment of this application, an electronic device is also provided, including: a memory and a processor. The processor is used to run a program stored in the memory, wherein the program executes the following method for scheduling discrete-action equipment in a distribution network: determining load forecast data of the distribution network and active power output forecast data of the distributed generation in the distribution network within a preset time period; calling a discrete-action equipment scheduling model to process the load forecast data and active power output forecast data, thereby obtaining a scheduling scheme for discrete-action equipment in the distribution network, wherein the scheduling scheme includes the predicted switching positions of the taps of the discrete-action equipment within the preset time period, the training data of the discrete-action equipment scheduling model includes labeled data and unlabeled data, the discrete-action equipment scheduling model includes a feature extractor for identifying the data input into the discrete-action equipment scheduling model, and an estimation module for generating the scheduling scheme; adjusting the tap switching positions of the discrete-action equipment within the preset time period according to the scheduling scheme.

[0084] According to an embodiment of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements the following method for scheduling discrete-operation equipment in a distribution network: determining load forecast data of the distribution network and active power output forecast data of the distributed generation in the distribution network within a preset time period; calling a discrete-operation equipment scheduling model to process the load forecast data and active power output forecast data, thereby obtaining a scheduling scheme for discrete-operation equipment in the distribution network, wherein the scheduling scheme includes the predicted switching positions of the taps of the discrete-operation equipment within the preset time period, the training data of the discrete-operation equipment scheduling model includes labeled data and unlabeled data, the discrete-operation equipment scheduling model includes a feature extractor for identifying the data input into the discrete-operation equipment scheduling model, and an estimation module for generating the scheduling scheme; and adjusting the tap switching positions of the discrete-operation equipment within the preset time period according to the scheduling scheme.

[0085] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0086] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0087] 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 units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0088] Furthermore, the functional units in the various embodiments of this application 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.

[0089] 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, the technical solution of this application, in essence, or the part that contributes to related technologies, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0090] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for scheduling discrete-action equipment in a power distribution network, characterized in that, include: Determine the load forecast data of the distribution network and the active power output forecast data of the distributed power sources of the distribution network within a preset time period. The load forecast data and the active power output forecast data are processed by calling the discrete motion equipment scheduling model to obtain the scheduling scheme of discrete motion equipment in the distribution network. The scheduling scheme includes the predicted switching positions of the taps of the discrete motion equipment. The training data of the discrete motion equipment scheduling model includes labeled data and unlabeled data. The discrete motion equipment scheduling model includes a feature extractor for identifying the data input into the discrete motion equipment scheduling model, and an estimation module for generating the scheduling scheme. The switching position of the tap changer of the discrete action device is adjusted within the preset time period according to the scheduling scheme; The discrete action device scheduling model is trained in the following way: Acquire historical data of the power distribution network, wherein the historical data includes historical load forecast data and historical active power output forecast data; Identify the labeled and unlabeled historical data within the historical data; The principal component features of the labeled historical data are determined, and the labeled historical data are filtered based on the principal component features to obtain the first training dataset; The principal component features of the labeled historical data are used as the reference principal component features of the unlabeled historical data, and the unlabeled historical data are clustered and filtered according to the reference principal component features to obtain the second training dataset. An adversarial generative network is used to train the discrete action device scheduling model based on the first training dataset and the second training dataset.

2. The method for scheduling discrete-action equipment in a distribution network according to claim 1, characterized in that, The discrete-action equipment scheduling model is invoked to process the load forecast data and the active power output forecast data, thereby obtaining the scheduling scheme for discrete-action equipment in the distribution network, including: The feature extractor in the discrete motion equipment scheduling model extracts data features from the load forecast data and the active power output forecast data, and performs data dimensionality reduction and global average pooling on the data features to obtain a data feature vector of a preset length. The estimation module in the device scheduling model determines the predicted location value of the data feature vector mapping through a fully connected layer, wherein the predicted location value is used to indicate the predicted switching position.

3. The method for scheduling discrete-action equipment in a distribution network according to claim 1, characterized in that, Determining the principal component features of the labeled historical data includes: The labeled data is standardized so that the mean of the labeled data is a first preset value and the variance is a second preset value. Determine the covariance matrix of the standardized labeled data, wherein the principal component features include principal component feature vectors; the covariance matrix is ​​used to reflect the interrelationships between the various feature quantities in the labeled data; The covariance matrix is ​​subjected to eigenvalue decomposition to obtain eigenvalues ​​and eigenvectors corresponding to the eigenvalues; The principal component feature vector in the labeled data is determined based on the magnitude of the feature values.

4. The method for scheduling discrete-action equipment in a distribution network according to claim 1, characterized in that, The labeled historical data is filtered based on the principal component features to obtain the first training dataset, which includes: Determine the correlation between the principal component characteristics and the node voltage characteristics of the distribution network; The labeled data is filtered based on the correlation to obtain the first training dataset.

5. The method for scheduling discrete-action equipment in a distribution network according to claim 4, characterized in that, Based on the reference principal component features, the unlabeled historical data is clustered and filtered to obtain the second training dataset, which includes: The unlabeled data is standardized so that the mean of the unlabeled data is a first preset value and the variance is a second preset value. A feature space is constructed based on the interrelationships reflected in the covariance matrix and the standardized unlabeled data; The principal component space is determined based on the reference principal components, and the feature space is processed in the principal component space to obtain the feature representation corresponding to the unlabeled historical data; Cluster analysis is performed on the feature representation to obtain the second training dataset. During the cluster analysis process, the influence of the tap position of the discrete motion device in the unlabeled data on the voltage control result is verified in terms of its manner and degree.

6. The method for scheduling discrete-action equipment in a distribution network according to claim 1, characterized in that, The method further includes: During training, the cross-entropy loss of the discrete action equipment scheduling model is determined, and the loss function value of the discrete action equipment scheduling model is determined based on the cross-entropy loss, as well as the gear switching penalty coefficient of the discrete action equipment in the distribution network. The gear switching penalty coefficient is used to adjust the loss function value, and the larger the gear switching penalty coefficient is, the larger the adjusted loss function value is.

7. A dispatching device for discrete action equipment in a power distribution network, characterized in that, include: The first processing module is used to determine the load forecast data of the distribution network within a preset time period, as well as the active power output forecast data of the distributed power sources of the distribution network. The second processing module is used to call the discrete motion equipment scheduling model to process the load forecast data and the active power output forecast data, thereby obtaining the scheduling scheme of discrete motion equipment in the distribution network. The scheduling scheme includes the predicted switching positions of the taps of the discrete motion equipment. The training data of the discrete motion equipment scheduling model includes labeled data and unlabeled data. The discrete motion equipment scheduling model includes a feature extractor for identifying the data input into the discrete motion equipment scheduling model, and an estimation module for generating the scheduling scheme. The third processing module is used to adjust the tap switching position of the discrete action device within the preset time period according to the scheduling scheme. The discrete action device scheduling model is trained in the following way: Acquire historical data of the power distribution network, wherein the historical data includes historical load forecast data and historical active power output forecast data; Identify the labeled and unlabeled historical data within the historical data; The principal component features of the labeled historical data are determined, and the labeled historical data are filtered based on the principal component features to obtain the first training dataset; The principal component features of the labeled historical data are used as the reference principal component features of the unlabeled historical data, and the unlabeled historical data are clustered and filtered according to the reference principal component features to obtain the second training dataset. An adversarial generative network is used to train the discrete action device scheduling model based on the first training dataset and the second training dataset.

8. A non-volatile storage medium, characterized in that, The non-volatile storage medium stores a program, wherein when the program is executed, it controls the device where the non-volatile storage medium is located to execute the power distribution network discrete action device scheduling method according to any one of claims 1 to 6.

9. An electronic device, characterized in that, include: A memory and a processor, the processor being configured to run a program stored in the memory, wherein the program, when running, executes the power distribution network discrete action device scheduling method according to any one of claims 1 to 6.

10. A computer program product, characterized in that, It includes a computer program that, when executed by a processor, implements the method for scheduling discrete action devices in a power distribution network according to any one of claims 1 to 6.

Images