Fault prediction method and system for multi-source heterogeneous power distribution network

Abstract

The invention provides a multi-source heterogeneous power distribution network fault prediction method and system, relates to the technical field of power distribution network fault identification, and aims to solve the problems that an existing model is single in consideration factor, cannot comprehensively simulate a complex fault type of a multi-source heterogeneous power distribution network, is large in calculation amount, is low in efficiency, and is poor in fault prediction accuracy. The method comprises the steps of obtaining historical data related to a power distribution fault; preprocessing the historical data related to the power distribution network fault to obtain preprocessed historical data; analyzing the preprocessed data by adopting a clustering technology to obtain a clustering result; based on a clustering result, constructing a fault prediction model for training to obtain a trained fault prediction model; and inputting real-time data related to the fault of the power distribution network into the trained fault prediction model for prediction. According to the method, the problems of delay, large calculation amount and low efficiency are avoided, the real operation condition of the power distribution network can be reflected, and the fault type of the multi-source heterogeneous power distribution network is comprehensively simulated.

Description

Technical Field

[0001] The present invention belongs to the technical field of distribution network fault identification, and in particular relates to a multi-source heterogeneous distribution network fault prediction method and system. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] The power system is an important part of the national economy. As a link in the power system, the distribution network directly distributes electricity to users. The safety and reliability of its power supply is related to the quality of users' electricity and may even have an impact on the national economy. The reliability assessment of the distribution network is extremely important.

[0004] Based on the current status of the distribution network, a complex distribution architecture with distributed, multi-load, and multi-power features is taking shape. The multi-source data of the distribution network is complex in type, non-uniform in structure, decentralized in storage, and poor in correlation. The requirements for power supply reliability are gradually increasing. However, in the event of a failure in the distribution network, power outages or accidents will cause significant economic losses and even threaten personal safety and affect social stability. Therefore, how to accurately and comprehensively predict multi-source heterogeneous distribution network failures is the key to the normal operation of the distribution network.

[0005] The inventors have found that the existing multi-source heterogeneous distribution network fault prediction technology has certain problems:

[0006] (1) The model considers only a few factors, cannot fully utilize operating data, cannot fully consider fault factors, and has poor reliability in fault detection;

[0007] (2) Before being put into use, a large number of experiments are needed to verify the reliability, stability and accuracy of the technology in different line topologies, different distribution network operation modes and various types of fault prediction. However, verification on the actual distribution network affects the safe operation of the power grid and consumes manpower and material resources. In addition, simulating faults through digital simulators cannot reflect the actual operation of the distribution network and cannot fully simulate the complex fault types of multi-source heterogeneous distribution networks. At the same time, when the data value is large, there will be delays when using instrument simulation, and the calculation amount is large and the efficiency is low, which in turn affects fault maintenance. Summary of the invention

[0008] In order to overcome the deficiencies of the above-mentioned prior art, the present invention provides a multi-source heterogeneous distribution network fault prediction method and system. By clustering the acquired multi-source heterogeneous distribution network data, the fusion of distribution network data is realized, providing more comprehensive data for prediction; the prediction of multi-source heterogeneous distribution network faults is realized through an improved neural network, which reduces the amount of calculation, improves efficiency and prediction accuracy, and reduces costs.

[0009] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions:

[0010] The first aspect of the present invention discloses a multi-source heterogeneous distribution network fault prediction method, comprising:

[0011] Obtain historical data related to power distribution faults, including distribution network data and weather data;

[0012] Preprocessing historical data related to distribution network faults to obtain preprocessed historical data;

[0013] Clustering technology is used to analyze the preprocessed data to obtain clustering results;

[0014] Based on the clustering results, a fault prediction model is constructed for training to obtain a trained fault prediction model;

[0015] The real-time data related to distribution network faults are input into the trained fault prediction model to obtain the prediction results.

[0016] As a further technical solution, the distribution network data includes distribution network time series data, data combining distribution network data with user-side network equipment data, and load data;

[0017] The distribution network time series data includes the current, voltage and power of each node of the distribution network at different times;

[0018] The user-side network device data includes handshake data and heartbeat data of the network device; the user-side network device includes a set-top box and a router.

[0019] As a further technical solution, the historical data related to the distribution network failure is preprocessed, specifically: the historical data related to the distribution network failure is cleaned to remove abnormal values ​​in the data;

[0020] The cleaned data is standardized to obtain the preprocessed historical data.

[0021] As a further technical solution, KFCM clustering technology is used to cluster the preprocessed data, specifically:

[0022] Determine the maximum number of fault categories and the initial number of distribution network fault clusters in the preprocessed historical data;

[0023] Calculate the initial cluster center and membership degree under the initial number of clusters of distribution network faults;

[0024] According to the clustering data set, the clustering objective function is constructed and iteratively updated;

[0025] When the iterative update reaches the minimum clustering objective function, the clustering division result is obtained.

[0026] As a further technical solution, the clustering objective function is formulated as follows:

[0027]

[0028] in, represents the i-th element in the fault clustering data, D j Indicates belonging to a i The jth element of ij is the jth membership degree corresponding to the i-th data, λ ij ∈[0,,1],

[0029] As a further technical solution, the fault prediction model adopts a YOLOv7 neural network, including an input layer, a hidden layer, and an output layer;

[0030] The processing process of the hidden layer is: activating the clustering data set through an activation function;

[0031] The sub-data of different categories in the input data are multiplied by their corresponding weight parameters and then added, and the added data and the bias term are summed to obtain fused data.

[0032] The second aspect discloses a multi-source heterogeneous distribution network fault prediction system, comprising:

[0033] A data acquisition module is used to acquire historical data related to power distribution faults, including distribution network data and weather data;

[0034] A data preprocessing module is used to preprocess historical data related to distribution network faults to obtain preprocessed historical data;

[0035] The cluster analysis module is used to analyze the pre-processed data using clustering technology to obtain clustering results;

[0036] The prediction model building and training module is used to build a fault prediction model based on the clustering results and conduct training to obtain a trained fault prediction model;

[0037] The prediction module is used to input the real-time data related to the distribution network fault into the trained fault prediction model to obtain the prediction result.

[0038] The third aspect of the present invention aims to provide a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps in the method described in the first aspect of the present invention.

[0039] A fourth aspect of the present invention provides a computer device, comprising a memory, a processor, and a program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps in the method described in the first aspect of the present invention are implemented.

[0040] A fifth aspect of the present invention provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute the methods and functions involved in any of the above-mentioned embodiments.

[0041] One or more of the above technical solutions have the following beneficial effects:

[0042] The prediction model in this embodiment can fully utilize the operation data, comprehensively consider the fault factors, and improve the reliability of model fault detection;

[0043] The prediction model in this embodiment can be trained, tested and verified before being put into use without affecting the safe operation of the power grid, saving time and effort. It can also reflect the actual operation of the distribution network, comprehensively simulate the complex fault types of multi-source heterogeneous distribution networks, avoid delays, large amounts of calculations, and low efficiency, and realize real-time and effective maintenance of faults.

[0044] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0046] Figure 1 This is a flow chart of a multi-source heterogeneous distribution network fault prediction method according to Embodiment 1 of the present invention. DETAILED DESCRIPTION

[0047] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0048] It should be noted that the terms used herein are for describing specific embodiments only and are not intended to be limiting of exemplary embodiments according to the present invention.

[0049] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0050] Embodiment 1

[0051] This embodiment discloses a multi-source heterogeneous distribution network fault prediction method.

[0052] In order to more clearly illustrate this embodiment, the multi-source heterogeneous distribution network fault prediction implementation process can be specifically described as follows:

[0053] A multi-source heterogeneous distribution network fault prediction method is provided, comprising:

[0054] S1. Obtain historical data related to power distribution faults, including distribution network data and weather data;

[0055] S2. Preprocessing the historical data related to the distribution network fault to obtain the preprocessed historical data;

[0056] S3, using clustering technology to analyze the preprocessed data to obtain clustering results;

[0057] S4. Based on the clustering results, a fault prediction model is constructed for training to obtain a trained fault prediction model;

[0058] S5. Input the real-time data related to the distribution network fault into the trained fault prediction model to obtain the prediction result.

[0059] like Figure 1 As shown, in step S1, historical data related to power distribution faults is obtained, and the historical data includes distribution network data and weather data.

[0060] In this embodiment, the distribution network data includes distribution network time series data, data obtained by combining distribution network data with user-side network equipment data, and load data.

[0061] The distribution network time series data includes the current, voltage and power of each node in the distribution network at different times.

[0062] User-side network device data includes handshake data and heartbeat data of network devices; user-side network devices include set-top boxes and routers.

[0063] In the embodiment, there is no need to establish an additional intelligent terminal. Without adding on-site hardware, accurate and effective historical data can be directly obtained from the power system to provide a data basis for subsequent model prediction.

[0064] like Figure 1 As shown, in step S2, historical data related to distribution network faults are preprocessed to obtain preprocessed historical data.

[0065] (1) Preprocessing the historical data related to distribution network failures, specifically: cleaning the historical data related to distribution network failures and removing abnormal values ​​in the data;

[0066] In this embodiment, due to the failure of distribution network sensors or data deviation and loss under harsh conditions, port abnormalities and input errors, etc., historical data errors, incomplete data or duplication are caused. This part of the data will affect the effect of model prediction, so data cleaning is performed first.

[0067] Redundant data is directly deleted, and missing values ​​are filled with default data values, mean filling, median filling, 0 filling, etc.

[0068] (2) The cleaned data is standardized to obtain preprocessed historical data.

[0069] The standardized formula is used for processing to obtain the preprocessed historical data.

[0070] like Figure 1 As shown, in step S3, clustering technology is used to analyze the preprocessed data to obtain clustering results.

[0071] In this embodiment, the KFCM clustering technology is used to cluster the preprocessed data, specifically:

[0072] (1) Determine the maximum number of fault categories and the initial number of distribution network fault clusters in the preprocessed historical data;

[0073] In this embodiment, the maximum number of fault categories Kmax and the initial number of clusters k in the historical data are determined.

[0074] (2) Calculate the initial cluster center and affiliation under the initial cluster number of distribution network faults.

[0075] F=(a 1 ,a 2 ,.......a k ) is the data set, D is the k clusters obtained after clustering, λ ij is the jth membership degree corresponding to the i-th data, λ ij ∈[0,,1], The initial cluster center is

[0076] (3) Based on the clustering data set, construct the clustering objective function and perform iterative updates;

[0077]

[0078] in, represents the i-th element in the fault clustering data, D j Indicates belonging to a i The jth element of ij is the j-th membership degree corresponding to the i-th data.

[0079] (4) When the iterative update reaches the minimum clustering objective function, the clustering division result is obtained.

[0080] After iterating the objective function, when the objective function is minimized, the clustering result is obtained.

[0081] like Figure 1 In step S4, based on the clustering result, a fault prediction model is constructed and trained to obtain a trained fault prediction model.

[0082] In this embodiment, the fault prediction model adopts the YOLOv7 neural network, which includes an input layer, a hidden layer, and an output layer.

[0083] The clustering results are input into the YOLOv7 neural network, and the time series data, weather data and load data in the historical data are used as different inputs respectively. The input layer of the neural network can receive these data respectively.

[0084] h=σ(Wx*X(t)+Ww*W(t)+W 1 *L(t)+b)(2)

[0085] Where b represents the bias term, σ represents the Sigmoid activation function), Wx, Ww, and Wl represent the weight parameters of the time series data, weather data, and load data to the hidden layer, respectively. The specific expression formula of Wx can be expressed as a matrix, in which each row corresponds to a neuron in the hidden layer, and each column corresponds to a feature of the time series data.

[0086] It is compared with the clustering results and passed through the YOLOv7 neural network to obtain the fault prediction results.

[0087] Embodiment 2

[0088] The purpose of this embodiment is to provide a multi-source heterogeneous distribution network fault prediction system, including:

[0089] A data acquisition module is used to acquire historical data related to power distribution faults, including distribution network data and weather data;

[0090] A data preprocessing module is used to preprocess historical data related to distribution network faults to obtain preprocessed historical data;

[0091] The cluster analysis module is used to analyze the pre-processed data using clustering technology to obtain clustering results;

[0092] The prediction model building and training module is used to build a fault prediction model based on the clustering results and conduct training to obtain a trained fault prediction model;

[0093] The prediction module is used to input the real-time data related to the distribution network fault into the trained fault prediction model to obtain the prediction result.

[0094] A multi-source heterogeneous distribution network fault prediction system in this embodiment executes a multi-source heterogeneous distribution network fault prediction method in the first embodiment.

[0095] Embodiment 3

[0096] The purpose of this embodiment is to provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the steps of the above method are implemented when the processor executes the program.

[0097] Embodiment 4

[0098] The purpose of this embodiment is to provide a computer-readable storage medium.

[0099] A computer-readable storage medium stores a computer program, which executes the steps of the above method when executed by a processor.

[0100] Embodiment 5

[0101] The purpose of this embodiment is to provide a computer program product containing instructions, which, when running on a computer, enables the computer to execute the methods and functions involved in any of the above embodiments.

[0102] The steps involved in the apparatus of the above embodiment correspond to the method embodiment 1, and the specific implementation method can refer to the relevant description part of embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood to include any medium that can store, encode or carry an instruction set for execution by a processor and enable the processor to execute any method in the present invention.

[0103] Those skilled in the art should understand that the modules or steps of the present invention described above can be implemented by a general-purpose computer device, or alternatively, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. The present invention is not limited to any specific combination of hardware and software.

[0104] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A multi-source heterogeneous distribution network fault prediction method, characterized in that: include: Obtain historical data related to power distribution faults, including distribution network data and weather data; Preprocessing historical data related to distribution network faults to obtain preprocessed historical data; Clustering technology is used to analyze the preprocessed data to obtain clustering results; Based on the clustering results, a fault prediction model is constructed for training to obtain a trained fault prediction model; The real-time data related to distribution network faults are input into the trained fault prediction model to obtain the prediction results.

2. A multi-source heterogeneous distribution network fault prediction method as claimed in claim 1, characterized in that: The distribution network data includes distribution network time series data, data combining distribution network data with user-side network equipment data, and load data; The distribution network time series data includes the current, voltage and power of each node of the distribution network at different times; The user-side network device data includes handshake data and heartbeat data of the network device; the user-side network device includes a set-top box and a router.

3. A multi-source heterogeneous distribution network fault prediction method as claimed in claim 1, characterized in that: Preprocess the historical data related to the distribution network failure, specifically: clean the historical data related to the distribution network failure and remove abnormal values ​​in the data; The cleaned data is standardized to obtain the preprocessed historical data.

4. A multi-source heterogeneous distribution network fault prediction method as claimed in claim 1, characterized in that: The KFCM clustering technology is used to cluster the preprocessed data, specifically: Determine the maximum number of fault categories and the initial number of distribution network fault clusters in the preprocessed historical data; Calculate the initial cluster center and membership degree under the initial number of clusters of distribution network faults; According to the clustering data set, the clustering objective function is constructed and iteratively updated; When the iterative update reaches the minimum clustering objective function, the clustering division result is obtained.

5. A multi-source heterogeneous distribution network fault prediction method as claimed in claim 4, characterized in that: in, represents the i-th element in the fault clustering data, D j Indicates belonging to a i The jth element of ij is the jth membership degree corresponding to the i-th data, λ ij ∈[0,,1], 6. A multi-source heterogeneous distribution network fault prediction method as claimed in claim 1, characterized in that: The fault prediction model adopts the YOLOv7 neural network, including an input layer, a hidden layer, and an output layer; The processing process of the hidden layer is: activating the clustering data set through an activation function; The sub-data of different categories in the input data are multiplied by their corresponding weight parameters and then added, and the added data and the bias term are summed to obtain fused data.

7. A multi-source heterogeneous distribution network fault prediction system, characterized in that: include: A data acquisition module is used to acquire historical data related to power distribution faults, including distribution network data and weather data; A data preprocessing module is used to preprocess historical data related to distribution network faults to obtain preprocessed historical data; The cluster analysis module is used to analyze the pre-processed data using clustering technology to obtain clustering results; The prediction model building and training module is used to build a fault prediction model based on the clustering results and conduct training to obtain a trained fault prediction model; The prediction module is used to input the real-time data related to the distribution network fault into the trained fault prediction model to obtain the prediction result.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the method described in any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method described in any one of claims 1 to 6 are performed.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images