Power distribution network state judgment method and system based on small sample state

Abstract

The embodiment of the invention provides a power distribution network state judgment method and system based on a small sample state, and belongs to the field of power system automation. The power distribution network state judgment method comprises the following steps: acquiring a sample set of a power distribution network state; constructing a training set according to the sample set; training according to the training set to obtain an enhanced sample set; constructing a power distribution network state judgment model according to the enhanced sample set; and judging the state of the power distribution network according to the state judgment model of the power distribution network. According to the power distribution network state judgment method and system, the accuracy and stability of power distribution network state judgment can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power system automation, and in particular to a distribution network state judgment method and system based on a small sample state. Background Art

[0002] With the development of society, the connection between power grids is becoming closer and more complex. At the same time, users have higher and higher requirements for power quality. Any unplanned power outage will cause immeasurable impact. Therefore, as the terminal connecting the power grid and users, it is very important to conduct real-time online monitoring of the operating status of the distribution network.

[0003] With the development of smart grids and the construction of strong power grids, the reliability of distribution network operation is constantly improving. A large part of the time during operation is in normal state, and faults rarely occur. The actual fault samples obtained through collection are relatively small. In order to improve the reliability of the distribution network, attention should be paid to the identification of its fault state. The massive normal state samples result in low data availability. Therefore, it is an extremely important topic to be able to quickly and accurately identify the online operation status of the distribution network under small fault samples. However, under the actual operation of the power grid, the data of the small fault samples obtained is single and the fault characteristics are not obvious, which makes it impossible to identify the operation status of the distribution network based on the existing data. Summary of the invention

[0004] The purpose of the embodiments of the present invention is to provide a method and system for determining the state of a distribution network based on a small sample state, wherein the method and system can improve the accuracy and stability of determining the state of the distribution network.

[0005] In order to achieve the above object, an embodiment of the present invention provides a method for determining a distribution network state based on a small sample state, comprising:

[0006] Obtain a sample set of distribution network status;

[0007] Constructing a training set according to the sample set;

[0008] Perform training according to the training set to obtain an enhanced sample set;

[0009] Constructing a distribution network state judgment model according to the enhanced sample set;

[0010] The state of the distribution network is judged according to the distribution network state judgment model.

[0011] Optionally, constructing a training set according to the sample set includes:

[0012] Extracting characteristic variables of state information in the sample set;

[0013] The characteristic variables are normalized to a set interval to obtain the training set.

[0014] Optionally, training is performed according to the training set to obtain an enhanced sample set, including:

[0015] Randomly initialize a generator of a conditional generative network and set the number of cycles of the generator;

[0016] Generate an experimental sample set according to the number of cycles with the minimum loss function as the objective function;

[0017] The generated samples with the largest geometric mean are selected from the experimental sample set and merged into the training set to obtain the enhanced sample set.

[0018] Optionally, generating an experimental sample set according to the number of cycles with a minimum loss function as an objective function includes:

[0019] Generate experimental samples;

[0020] According to the experimental samples and the training set, the accuracy rate is judged by the discriminator;

[0021] The parameters of the generator are modified according to the discrimination result.

[0022] Optionally, selecting a generated sample with the largest geometric mean from the experimental sample set and fusing it into the training set to obtain the enhanced sample set includes:

[0023] According to the experimental sample set, the generated sample with the largest geometric mean is calculated according to formula (1):

[0024]

[0025] Among them, p1 to p6 are respectively the classification recognition accuracy rates of the generated samples under the original classification model for the six operating states of the distribution network.

[0026] Optionally, constructing a distribution network state judgment model according to the enhanced sample set includes:

[0027] Normalizing the feature values ​​of the enhanced sample set within a preset interval to obtain normalized feature values, and initializing a neural network;

[0028] Training the neural network according to the normalized eigenvalues ​​to obtain a network error;

[0029] Determine whether the network error is less than a preset target error or the number of training times is greater than or equal to a preset number threshold;

[0030] In the case where it is determined that the network error is less than the target error or the number of training times is greater than or equal to the number threshold, obtaining the distribution network state judgment model;

[0031] When it is determined that the network error is greater than or equal to the target error or the number of training times is less than the number threshold, continue training.

[0032] On the other hand, the present invention further provides a distribution network state judgment system based on small sample states, the distribution network state judgment system comprising a controller, and the controller is used to execute any of the distribution network state judgment methods described above.

[0033] Through the above technical scheme, the distribution network state judgment method and system based on small sample state provided by the present invention enhance the acquired sample set through conditional generative adversarial network, generate more realistic data through nonlinear interpolation, and integrate the generated data set into the original sample set to synthesize sufficient data volume and more realistic training samples for training to obtain the distribution network state judgment model and then judge the distribution network state, thereby improving the accuracy and stability of the distribution network state judgment.

[0034] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings:

[0036] Figure 1 is a flow chart of a method for determining a state of a distribution network according to an embodiment of the present invention;

[0037] Figure 2 is a flowchart of a method for constructing a training set based on a sample set in a method for determining a distribution network state according to an embodiment of the present invention;

[0038] Figure 3 is a flowchart of a method for obtaining an enhanced sample set by training according to a training set in a method for determining a distribution network state according to an embodiment of the present invention;

[0039] Figure 4 It is a flowchart of generating an experimental sample set according to the number of cycles with the minimum loss function as the objective function in a method for judging the state of a distribution network according to an embodiment of the present invention;

[0040] Figure 5The present invention is a flowchart of a method for constructing a distribution network state judgment model according to an enhanced sample set in a distribution network state judgment method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0041] The specific implementation of the embodiment of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiment of the present invention, and is not used to limit the embodiment of the present invention.

[0042] It should be noted that the acquisition, transmission, storage, use, and processing of data in the technical solution of this application are in compliance with the relevant provisions of national laws and regulations. In the embodiments of this application, some existing solutions in the industry such as certain software, components, and models may be mentioned, which should be considered as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of this application, but it does not mean that the applicant has or will necessarily use the solution.

[0043] like Figure 1 FIG. 1 is a flow chart of a method for determining a distribution network state according to an embodiment of the present invention. Figure 1 The method for determining the state of the distribution network may include the following steps:

[0044] In step S10, a sample set of distribution network status is obtained;

[0045] In step S11, a training set is constructed according to the sample set;

[0046] In step S12, training is performed according to the training set to obtain an enhanced sample set;

[0047] In step S13, a distribution network state judgment model is constructed according to the enhanced sample set;

[0048] In step S14, the state of the distribution network is judged according to the distribution network state judgment model.

[0049] In this Figure 1 In the method shown, step S11 is used to construct a training set based on the sample set. In this embodiment, the method for constructing a training set based on the sample set can be in various forms known to those skilled in the art. In a preferred example of the present invention, the method for constructing a training set based on the sample set can include: Figure 2 The steps shown. Figure 2 In the method of constructing a training set according to a sample set, the method may include the following steps:

[0050] In step S110, feature variables of state information in the sample set are extracted;

[0051] In step S111, the feature variables are normalized to a set interval to obtain a training set.

[0052] In this Figure 1 In the method shown, step S12 is used to perform training according to the training set to obtain an enhanced sample set. In this embodiment, the method for performing training according to the training set to obtain an enhanced sample set can be in various forms known to those skilled in the art. In a preferred example of the present invention, the method for performing training according to the training set to obtain an enhanced sample set can include: Figure 3 The steps shown. Figure 3 In the method of training according to the training set to obtain the enhanced sample set, the method may include the following steps:

[0053] In step S120, the generator of the conditional generative network is randomly initialized and the number of cycles of the generator is set;

[0054] In step S121, an experimental sample set is generated according to the number of cycles with the minimum loss function as the objective function;

[0055] In step S122, the generated samples with the largest geometric mean are selected from the experimental sample set and merged into the training set to obtain an enhanced sample set.

[0056] In this Figure 3 In the method shown, step S121 is used to generate an experimental sample set according to the number of cycles with the minimum loss function as the objective function. In this embodiment, the method for generating an experimental sample set according to the number of cycles with the minimum loss function as the objective function can be in various forms known to those skilled in the art. In a preferred example of the present invention, generating an experimental sample set according to the number of cycles with the minimum loss function as the objective function can include: Figure 4 The steps shown. Figure 4 In the process, generating an experimental sample set based on the number of cycles and taking the minimum loss function as the objective function may include the following steps:

[0057] In step S1210, an experimental sample is generated;

[0058] In step S1211, the accuracy is judged by a discriminator according to the experimental samples and the training set;

[0059] In step S1212, the parameters of the generator are modified according to the discrimination result.

[0060] In this Figure 3In the method shown, step S122 is used to select the generated sample with the largest geometric mean according to the experimental sample set and merge it into the training set to obtain an enhanced sample set. In this embodiment, the method for selecting the generated sample with the largest geometric mean according to the experimental sample set and merging it into the training set to obtain an enhanced sample set can be in various forms known to those skilled in the art. In a preferred example of the present invention, the method for selecting the generated sample with the largest geometric mean according to the experimental sample set and merging it into the training set to obtain an enhanced sample set may include:

[0061] According to the experimental sample set, the generated sample with the largest geometric mean is calculated according to formula (1):

[0062]

[0063] Among them, p1 to p6 are the classification and recognition accuracy of the generated samples under the original classification model for the six operating states of the distribution network.

[0064] In this Figure 1 In the method shown, step 13 is used to construct a distribution network state judgment model based on the enhanced sample set. In this embodiment, the method for constructing a distribution network state judgment model based on the enhanced sample set can be in various forms known to those skilled in the art. In a preferred example of the present invention, the method for constructing a distribution network state judgment model based on the enhanced sample set can include: Figure 5 The steps shown. Figure 5 In the method for constructing a distribution network state judgment model according to an enhanced sample set, the method may include the following steps:

[0065] In step S130, the feature values ​​of the enhanced sample set are normalized within a preset interval to obtain normalized feature values, and the neural network is initialized;

[0066] In step S131, the neural network is trained according to the normalized eigenvalues ​​to obtain a network error;

[0067] In step S132, it is determined whether the network error is less than a preset target error or the number of training times is greater than or equal to a preset number threshold;

[0068] In step S133, when it is judged that the network error is less than the target error or the number of training times is greater than or equal to the number threshold, the distribution network state judgment model is obtained; when it is judged that the network error is greater than or equal to the target error or the number of training times is less than the number threshold, the training is continued, and the step of training the neural network according to the normalized eigenvalue to obtain the network error is returned, that is, the execution of step S131 is returned.

[0069] On the other hand, the present invention further provides a distribution network state judgment system based on small sample states, the distribution network state judgment system comprising a controller, and the controller is used to execute any of the distribution network state judgment methods described above.

[0070] Through the above technical scheme, the distribution network state judgment method and system based on small sample state provided by the present invention enhance the acquired sample set through conditional generative adversarial network, generate more realistic data through nonlinear interpolation, and integrate the generated data set into the original sample set to synthesize sufficient data volume and more realistic training samples for training to obtain the distribution network state judgment model and then judge the distribution network state, thereby improving the accuracy and stability of the distribution network state judgment.

[0071] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0072] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0073] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0074] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0075] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0076] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0077] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0078] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0079] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A distribution network state judgment method based on small sample state, characterized in that: The distribution network state judgment method comprises: Obtain a sample set of distribution network status; Constructing a training set according to the sample set; Perform training according to the training set to obtain an enhanced sample set; Constructing a distribution network state judgment model according to the enhanced sample set; The state of the distribution network is judged according to the distribution network state judgment model.

2. The method for determining the state of a distribution network according to claim 1, characterized in that: Constructing a training set according to the sample set includes: Extracting characteristic variables of state information in the sample set; The characteristic variables are normalized to a set interval to obtain the training set.

3. The method for determining the state of a distribution network according to claim 1, characterized in that: Training is performed according to the training set to obtain an enhanced sample set, including: Randomly initialize a generator of a conditional generative network and set the number of cycles of the generator; Generate an experimental sample set according to the number of cycles with the minimum loss function as the objective function; The generated samples with the largest geometric mean are selected from the experimental sample set and merged into the training set to obtain the enhanced sample set.

4. The method for determining the state of a distribution network according to claim 3, characterized in that: Generate an experimental sample set according to the number of cycles with the minimum loss function as the objective function, including: Generate experimental samples; According to the experimental samples and the training set, the accuracy rate is judged by the discriminator; The parameters of the generator are modified according to the discrimination result.

5. The method for determining the state of a distribution network according to claim 3, characterized in that: The generated sample with the largest geometric mean is selected from the experimental sample set and merged into the training set to obtain the enhanced sample set, including: According to the experimental sample set, the generated sample with the largest geometric mean is calculated according to formula (1): Among them, p1 to p6 are respectively the classification recognition accuracy rates of the generated samples under the original classification model for the six operating states of the distribution network.

6. The method for determining the state of a distribution network according to claim 1, characterized in that: Constructing a distribution network state judgment model according to the enhanced sample set includes: Normalizing the feature values ​​of the enhanced sample set within a preset interval to obtain normalized feature values, and initializing a neural network; Training the neural network according to the normalized eigenvalues ​​to obtain a network error; Determine whether the network error is less than a preset target error or the number of training times is greater than or equal to a preset number threshold; In the case where it is determined that the network error is less than the target error or the number of training times is greater than or equal to the number threshold, obtaining the distribution network state judgment model; When it is determined that the network error is greater than or equal to the target error or the number of training times is less than the number threshold, continue training.

7. A distribution network status judgment system based on small sample status, characterized in that: The distribution network state judgment system comprises a controller, and the controller is used to execute the distribution network state judgment method according to any one of claims 1 to 6.

Images