CN113505876A - High-voltage circuit breaker fault diagnosis method based on generation type countermeasure network - Google Patents

High-voltage circuit breaker fault diagnosis method based on generation type countermeasure network Download PDF

Info

Publication number
CN113505876A
CN113505876A CN202110654484.1A CN202110654484A CN113505876A CN 113505876 A CN113505876 A CN 113505876A CN 202110654484 A CN202110654484 A CN 202110654484A CN 113505876 A CN113505876 A CN 113505876A
Authority
CN
China
Prior art keywords
data
fault
model
generative
voltage circuit
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202110654484.1A
Other languages
Chinese (zh)
Inventor
冯跃亮
张喆
蔡奎
孙献春
陈冰晶
宿波
周刚
过浩
沈云超
朱奕琦
周子誉
陈刚国
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jiaxing Hengchuang Electric Power Group Co ltd Bochuang Material Branch
Jiaxing Power Supply Co of State Grid Zhejiang Electric Power Co Ltd
Original Assignee
Jiaxing Hengchuang Electric Power Group Co ltd Bochuang Material Branch
Jiaxing Power Supply Co of State Grid Zhejiang Electric Power Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Jiaxing Hengchuang Electric Power Group Co ltd Bochuang Material Branch, Jiaxing Power Supply Co of State Grid Zhejiang Electric Power Co Ltd filed Critical Jiaxing Hengchuang Electric Power Group Co ltd Bochuang Material Branch
Priority to CN202110654484.1A priority Critical patent/CN113505876A/en
Publication of CN113505876A publication Critical patent/CN113505876A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

本发明公开了一种基于生成式对抗网络的高压断路器故障诊断方法，包括以下步骤：采集实际故障数据并进行数据预处理；构建生成式对抗网络模型；将随机噪声输入生成模型中，训练得到生成数据；将实际故障数据和生成数据输入判别模型进行真伪判别；输入数据训练生成式对抗网络，直到达到设定判别阈值；将实际故障数据和生成数据融合构成故障数据库。上述技术方案通过使用生成式对抗网络生成故障数据，丰富故障数据库，并且同时对生成的故障数据进行鉴别，判断数据是否异常、缺失，从而保证数据库数据的高质量，满足基于机器学习的高压断路器故障识别对高质量数据的要求，大大增加诊断结果的准确性。

Figure 202110654484

The invention discloses a fault diagnosis method for a high-voltage circuit breaker based on a generative confrontation network, comprising the following steps: collecting actual fault data and performing data preprocessing; building a generative confrontation network model; Generate data; input the actual fault data and generated data into the discrimination model for authenticity discrimination; train the generative adversarial network with the input data until the set discrimination threshold is reached; fuse the actual fault data and the generated data to form a fault database. The above technical solution generates fault data by using a generative confrontation network, enriches the fault database, and at the same time discriminates the generated fault data to determine whether the data is abnormal or missing, so as to ensure the high quality of the database data and meet the requirements of high-voltage circuit breakers based on machine learning. The requirement of high-quality data for fault identification greatly increases the accuracy of diagnostic results.

Figure 202110654484

Description

High-voltage circuit breaker fault diagnosis method based on generation type countermeasure network
Technical Field
The invention relates to the technical field of breaker diagnosis, in particular to a high-voltage breaker fault diagnosis method based on a generative countermeasure network.
Background
It is well known that circuit breakers are important devices in high voltage networks, mainly for controlling and protecting the devices. With the increase of the service time, certain parts of the circuit breaker can be continuously aged or damaged, such as the abrasion of a mechanical mechanism, the abrasion of a contact, the failure of a secondary opening and closing mechanism and the like. These problems can cause the circuit breaker to malfunction and even affect the safe operation of the entire grid. Therefore, it is important to periodically check the performance of the circuit breaker and evaluate its operating state.
The high-voltage circuit breaker is monitored in real time and diagnosed in fault, the running state of the circuit breaker can be grasped in real time, a reasonable maintenance plan is formulated according to the condition, and the method has important significance for guaranteeing safe running of a power grid. According to the IEEE recommendations for selecting breaker monitoring objects, the most common state quantities are: the circuit breaker comprises a circuit breaker moving contact stroke, a circuit breaker contact and ambient temperature, an opening and closing state, an opening and closing coil current, an energy storage motor current and circuit breaker mechanical vibration. The fault diagnosis of the common circuit breaker can be completed by collecting, feature extracting and classifying the parameters. In actual operation, however, fault data is extremely difficult to obtain. The circuit breaker simulation fault experiment can obtain high-quality data, but the experiment cost is higher; the simulation software can acquire fault simulation data, but the actual value and the simulation value have obvious errors, so that the acquisition of accurate data for realizing diagnosis becomes important.
Chinese patent document CN108828441A discloses a "fault diagnosis method for high-voltage circuit breaker". Adopts the following steps: 1) collecting the current of a switching-on and switching-off coil of a high-voltage circuit breaker to form a data set; 2) coding is carried out aiming at different state types of the related circuit breaker, data are randomly extracted according to a proportion to be respectively used as a pre-training sample, a parameter fine-tuning sample and a test sample, and normalization processing is carried out; 3) establishing a DBN classification network model, respectively using unlabeled pre-training data samples to perform network pre-training, and using labeled samples to perform parameter fine-tuning; 4) and (4) carrying out fault diagnosis on the high-voltage circuit breaker by using the trained DBN. According to the technical scheme, the fault simulation data can be obtained, but obvious errors exist between the fault simulation data and an actual fault value, and the accuracy of a diagnosis result is difficult to guarantee.
Disclosure of Invention
The invention mainly solves the technical problems that the fault simulation data acquired by the original technical scheme have obvious errors with the actual fault value and the fault diagnosis is realized due to lack of accurate data, and provides a high-voltage circuit breaker fault diagnosis method based on a generative countermeasure network.
The technical problem of the invention is mainly solved by the following technical scheme: the invention comprises the following steps:
s1, collecting actual fault data and carrying out data preprocessing;
s2, constructing a generative confrontation network model;
s3, inputting random noise into the generation model, and training to obtain generation data;
s4, inputting the actual fault data and the generated data into a discrimination model for discrimination;
s5, inputting data to train a generative confrontation network until a set discrimination threshold is reached;
s6 merges the actual fault data and the generated data to form a fault database.
Preferably, the step S1 of preprocessing the data includes integrating and deleting missing and repeated data, performing normalization processing and feature selection on the data set, and implementing classification of fault data, so as to accelerate the convergence speed of the sample generation model and reduce the data dimension.
Preferably, the generative countermeasure network model of step S2 includes a generative model for generating data and a discriminant model for determining whether the generated data is true or false and generating a data type, and the generative model and the discriminant model are mutually game-learned and are mutually exchanged and compared with the data of the fault database.
Preferably, the generated model is formed by a depth transposition convolution network, and the fault sample generation model is trained through the sampled real fault samples to learn the distribution of the real fault samples.
Preferably, the discrimination model is composed of a deep neural network, and the discrimination model includes a classifier for determining authenticity of data. And training an evaluation model by using the real fault sample to select the generated sample, wherein the generated fault sample used for training the fault diagnosis model is close to the distribution of the real sample so as to evaluate the quality of the generated sample.
Preferably, in step S4, the difference between the real fault data and the generated fault data is identified and distinguished by using a two-stage authenticity classifier, the output value is a binary number 0 or 1, if the output value is 0, the data is false data, if the output value is 1, the data is true data, and the generation model and the discriminant model are trained by using a large amount of data, so that the output value of the discriminant model after training is closer to 1.
Preferably, the first layer of the generated model is a fully-connected layer with a size of 6144, and the subsequent four layers are three-dimensional transposed convolutional layers, wherein the size of a convolutional kernel is 2 × 5 × 5, namely, time × width × height, and the transposed convolutional layer operation is used in the generated model. The use of transposed convolutional layer operations is considered to be a "reverse" propagation process of conventional convolutional operations.
Preferably, the discriminant model uses a convolutional neural network, the first four layers of the discriminant model are three-dimensional convolutional layers for extracting spatio-temporal features of data, followed by fully-connected layers, and the convolutional layers have a convolutional kernel size of 2 × 5 × 5, i.e., time × width × height.
Preferably, in step S3, the Wasserstein distance calculation formula is defined as the loss function of the Wasserstein distance optimization composite network model to stabilize the training process
Figure BDA0003113255020000041
Wherein, T1For distribution of true fault data obeys, T2Is the distribution, II (T), to which the generated data obeys1，T2) Is T1And T2A set of all joint distributions combined, γ being one of the joint distributions, (x, y) being a set of samples in γ, E(x，y)～γ[‖x-y‖]Is the expected value of the sample distance.
Preferably, in step S5, if it is determined that the true probability of the set of input data is greater than the set threshold, the set of input data is determined to be valid. The training aims to enable the discrimination truth rate of the generated confrontation network model to reach a set threshold value, alternately optimize and learn the generated model and the discrimination model, and finally achieve the zero-sum game.
The invention has the beneficial effects that: the fault data are generated by using the generative countermeasure network, so that a fault database is enriched, and the generated fault data are identified at the same time to judge whether the data are abnormal or missing, so that the high quality of the database data is ensured, the requirement of machine learning-based high-voltage circuit breaker fault identification on the high-quality data is met, and the accuracy of a diagnosis result is greatly improved.
Drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
The technical scheme of the invention is further specifically described by the following embodiments and the accompanying drawings. Example (b): the method for diagnosing the fault of the high-voltage circuit breaker based on the generative countermeasure network in the embodiment is shown in fig. 1, and comprises the following steps:
s1, collecting actual fault data and preprocessing the data, integrating and deleting missing and repeated data, and performing normalization processing and feature selection on a data set to realize classification of the fault data, so that the convergence speed of a sample generation model is increased, and the data dimensionality is reduced.
S2, constructing a generative confrontation network model, wherein the generative confrontation network model comprises a generative model for generating data and a discriminant model for judging whether the generated data is true or false and generating data categories, and the generative model and the discriminant model are mutually game-learned and are mutually exchanged and compared with the data of the fault database. The generation model is composed of a depth transposition convolution network, and the distribution of real fault samples is learned by training the fault sample generation model through the sampled real fault samples. The first layer of the generated model is a fully connected layer of 6144 size, the subsequent four layers are three-dimensional transposed convolutional layers, where the convolutional kernel size is 2 × 5 × 5, i.e., time × width × height, and the transposed convolutional layers are used in the generated model. The use of transposed convolutional layer operations is considered to be a "reverse" propagation process of conventional convolutional operations.
The discrimination model is composed of a deep neural network, and the discrimination model comprises a classifier for judging the authenticity of data. And training an evaluation model by using the real fault sample to select the generated sample, wherein the generated fault sample used for training the fault diagnosis model is close to the distribution of the real sample so as to evaluate the quality of the generated sample. The discriminative model uses a convolutional neural network, the first four layers of which are three-dimensional convolutional layers for extracting spatio-temporal features of data, followed by fully-connected layers, the convolutional layers having convolutional kernel sizes of 2 × 5 × 5, i.e., time × width × height.
S3, inputting random noise into the generation model, and training to obtain generation data. Using Wasserstein distance to optimize the loss function of the composite network model to stabilize the training process, and defining Wasserstein distance calculation formula as
Figure BDA0003113255020000051
Wherein, T1For distribution of true fault data obeys, T2Is the distribution, II (T), to which the generated data obeys1，T2) Is T1And T2A set of all joint distributions combined, γ being one of the joint distributions, (x, y) being a set of samples in γ, E(x，y)～γ[‖x-y‖]Is the expected value of the sample distance.
S4 inputs the actual failure data and the generated data to the discrimination model to discriminate the authenticity. And (3) identifying and distinguishing the difference between the real fault data and the generated fault data by using a true-false two-classifier, wherein the output value is binary number 0 or 1, if the output value is 0, the data is false data, if the output value is 1, the data is true data, and a large amount of data is used for training the generation model and the discrimination model, so that the output value of the trained discrimination model is closer to 1.
And S5, training the generative confrontation network by the input data until reaching a set discrimination threshold, and if the true probability of the set of input data is judged to be greater than the set threshold, judging the set of input data to be valid. The training aims to enable the discrimination truth rate of the generated confrontation network model to reach a set threshold value, alternately optimize and learn the generated model and the discrimination model, and finally achieve the zero-sum game.
S6 merges the actual fault data and the generated data to form a fault database.
The specific embodiments described herein are merely illustrative of the spirit of the invention. Various modifications or additions may be made to the described embodiments or alternatives may be employed by those skilled in the art without departing from the spirit or ambit of the invention as defined in the appended claims.
Although terms such as generative confrontation network models are used more herein, the possibility of using other terms is not excluded. These terms are used merely to more conveniently describe and explain the nature of the present invention; they are to be construed as being without limitation to any additional limitations that may be imposed by the spirit of the present invention.

Claims (10)

1.一种基于生成式对抗网络的高压断路器故障诊断方法，其特征在于，包括以下步骤：1. a high-voltage circuit breaker fault diagnosis method based on generative confrontation network, is characterized in that, comprises the following steps: S1采集实际故障数据并进行数据预处理；S1 collects actual fault data and performs data preprocessing; S2构建生成式对抗网络模型；S2 builds a generative adversarial network model; S3将随机噪声输入生成模型中，训练得到生成数据；S3 inputs random noise into the generation model, and trains to obtain the generated data; S4将实际故障数据和生成数据输入判别模型进行真伪判别；S4 inputs the actual fault data and the generated data into the discrimination model for authenticity discrimination; S5输入数据训练生成式对抗网络，直到达到设定判别阈值；S5 input data to train the generative adversarial network until the set discrimination threshold is reached; S6将实际故障数据和生成数据融合构成故障数据库。S6 fuses the actual fault data and the generated data to form a fault database. 2.根据权利要求1所述的一种基于生成式对抗网络的高压断路器故障诊断方法，其特征在于，所述步骤S1数据预处理包括通过对将缺失和重复的数据整合删除，将数据集进行归一化处理和特征选取，实现故障数据的分类，以加快样本生成模型的收敛速度，减少数据维度。2 . The method for diagnosing high-voltage circuit breakers based on generative adversarial networks according to claim 1 , wherein the data preprocessing in step S1 includes integrating and deleting missing and duplicated data, and synthesizing the data set. 3 . Perform normalization processing and feature selection to realize the classification of fault data, so as to speed up the convergence speed of the sample generation model and reduce the data dimension. 3.根据权利要求1所述的一种基于生成式对抗网络的高压断路器故障诊断方法，其特征在于，所述步骤S2生成式对抗网络模型包括用于生成数据的生成模型和用于判断生成数据真假以及生成数据类别的判别模型，生成模型与判别模型相互博弈学习并且与故障数据库的数据进行相互交换对比。3. A kind of high-voltage circuit breaker fault diagnosis method based on generative adversarial network according to claim 1, is characterized in that, described step S2 generative adversarial network model comprises the generation model that is used to generate data and is used to judge generation The data is true and false and the discriminant model of the generated data category. The generative model and the discriminant model learn from each other and exchange and compare with the data of the fault database. 4.根据权利要求3所述的一种基于生成式对抗网络的高压断路器故障诊断方法，其特征在于，所述生成模型由深度转置卷积网络构成，通过采样到的真实故障样本训练故障样本生成模型，学习真实故障样本的分布。4. The method for diagnosing faults of high-voltage circuit breakers based on generative adversarial networks according to claim 3, wherein the generative model is composed of a deep transposed convolutional network, and the faults are trained by sampling real fault samples The sample generation model learns the distribution of real fault samples. 5.根据权利要求1所述的一种基于生成式对抗网络的高压断路器故障诊断方法，其特征在于，所述判别模型由深度神经网络构成，所述判别模型中包含用于判断数据真伪的分类器。5. The method for diagnosing faults of high-voltage circuit breakers based on generative adversarial networks according to claim 1, wherein the discriminant model is composed of a deep neural network, and the discriminant model includes a method for judging the authenticity of the data. classifier. 6.根据权利要求1所述的一种基于生成式对抗网络的高压断路器故障诊断方法，其特征在于，所述步骤S4使用真伪二分类器鉴别区分真实故障数据和生成故障数据的差异，输出值为二进制数0或者1,若输出为0，则数据为假数据，若输出为1，则数据为真数据，使用大量数据训练生成模型和判别模型，使得训练后的判别模型输出值更接近于1。6. a kind of high-voltage circuit breaker fault diagnosis method based on generative adversarial network according to claim 1, is characterized in that, described step S4 uses true and false two-classifier to distinguish the difference between real fault data and generation fault data, The output value is a binary number of 0 or 1. If the output is 0, the data is false data, and if the output is 1, the data is true data. Use a large amount of data to train the generative model and the discriminant model, so that the output value of the discriminant model after training is more accurate. close to 1. 7.根据权利要求4所述的一种基于生成式对抗网络的高压断路器故障诊断方法，其特征在于，所述生成模型的第一层为6144大小的全连接层，后续四层是三维的转置卷积层，其中卷积核大小为2×5×5，即时间×宽×高，所述生成模型中使用转置卷积层操作。7. The method for diagnosing faults of high-voltage circuit breakers based on generative adversarial networks according to claim 4, wherein the first layer of the generative model is a fully connected layer with a size of 6144, and the subsequent four layers are three-dimensional The transposed convolutional layer, wherein the size of the convolution kernel is 2×5×5, that is, time×width×height, and the transposed convolutional layer operation is used in the generative model. 8.根据权利要求5所述的一种基于生成式对抗网络的高压断路器故障诊断方法，其特征在于，所述判别模型使用卷积神经网络，所述判别模型的前四层是用于提取数据的时空特征的三维卷积层，随后是全连接层，所述卷积层的卷积核大小为2×5×5，即时间×宽×高。8. The method for diagnosing faults of high-voltage circuit breakers based on generative adversarial networks according to claim 5, wherein the discriminant model uses a convolutional neural network, and the first four layers of the discriminant model are used to extract A three-dimensional convolutional layer of spatiotemporal features of the data, followed by a fully connected layer, has a convolutional kernel size of 2×5×5, ie time×width×height. 9.根据权利要求1所述的一种基于生成式对抗网络的高压断路器故障诊断方法，其特征在于，所述步骤S3使用Wasserstein距离优化复合网络模型的损失函数以稳定训练过程，Wasserstein距离计算式定义为9. a kind of high-voltage circuit breaker fault diagnosis method based on generative confrontation network according to claim 1, is characterized in that, described step S3 uses the loss function of Wasserstein distance optimization composite network model to stabilize training process, Wasserstein distance calculation is defined as
Figure FDA0003113255010000021
Figure FDA0003113255010000021
 其中，T1为真实故障数据服从的分布，T2是生成数据服从的分布，∏(T1，T2)为T1和T2组合起来的所有联合分布的集合，γ为其中的一个联合分布，(x，y)为γ中的一组样本，E(x，y)～γ[‖x-y‖]为该样本距离的期望值。Among them, T 1 is the distribution obeyed by the real fault data, T 2 is the distribution obeyed by the generated data, ∏(T 1 , T 2 ) is the set of all joint distributions combined by T 1 and T 2 , and γ is one of the joint distributions. distribution, (x, y) is a set of samples in γ, E (x, y) ~ γ [‖xy‖] is the expected value of the sample distance. 10.根据权利要求1所述的一种基于生成式对抗网络的高压断路器故障诊断方法，其特征在于，所述步骤S5若判别该组输入数据的真实概率大于设定阈值，则该组输入数据判定为有效。10. The method for diagnosing high-voltage circuit breakers based on generative adversarial networks according to claim 1, wherein in step S5, if it is determined that the true probability of the group of input data is greater than the set threshold, the group of input data The data is judged to be valid.
CN202110654484.1A 2021-06-11 2021-06-11 High-voltage circuit breaker fault diagnosis method based on generation type countermeasure network Pending CN113505876A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110654484.1A CN113505876A (en) 2021-06-11 2021-06-11 High-voltage circuit breaker fault diagnosis method based on generation type countermeasure network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110654484.1A CN113505876A (en) 2021-06-11 2021-06-11 High-voltage circuit breaker fault diagnosis method based on generation type countermeasure network

Publications (1)

Publication Number Publication Date
CN113505876A true CN113505876A (en) 2021-10-15

Family

ID=78010009

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110654484.1A Pending CN113505876A (en) 2021-06-11 2021-06-11 High-voltage circuit breaker fault diagnosis method based on generation type countermeasure network

Country Status (1)

Country Link
CN (1) CN113505876A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113900860A (en) * 2021-10-27 2022-01-07 重庆邮电大学 CGAN-based data recovery method for wireless sensor network fault node
CN115169506A (en) * 2022-09-06 2022-10-11 中铁第四勘察设计院集团有限公司 Method and system for rapidly diagnosing faults of power supply and transformation key equipment
CN116429406A (en) * 2023-06-14 2023-07-14 山东能源数智云科技有限公司 Construction method and device of fault diagnosis model of large-scale mechanical equipment

Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108680807A (en) * 2018-05-17 2018-10-19 国网山东省电力公司青岛供电公司 The Diagnosis Method of Transformer Faults and system of network are fought based on condition production
CN108805418A (en) * 2018-05-22 2018-11-13 福州大学 A kind of traffic data fill method fighting network based on production
JP2019056975A (en) * 2017-09-19 2019-04-11 株式会社Ｐｒｅｆｅｒｒｅｄ Ｎｅｔｗｏｒｋｓ Improved generative adversarial network achievement program, improved generative adversarial network achievement device, and learned model generation method
CN109813542A (en) * 2019-03-15 2019-05-28 中国计量大学 Fault Diagnosis Method of Air Handling Unit Based on Generative Adversarial Network
CN109918999A (en) * 2019-01-22 2019-06-21 西安交通大学 An intelligent diagnosis method for mechanical equipment faults based on generative model under small sample data
US20190197358A1 (en) * 2017-12-21 2019-06-27 International Business Machines Corporation Generative Adversarial Network Medical Image Generation for Training of a Classifier
CN110955651A (en) * 2019-11-28 2020-04-03 武汉科技大学 Motor fault data enhancement method based on deep convolution generation type countermeasure network
US20200202221A1 (en) * 2018-12-20 2020-06-25 Shandong University Of Science And Technology Fault detection method and system based on generative adversarial network and computer program
CN111445147A (en) * 2020-03-27 2020-07-24 中北大学 Generative confrontation network model evaluation method for mechanical fault diagnosis
CN111582348A (en) * 2020-04-29 2020-08-25 武汉轻工大学 Method, device, equipment and storage medium for training condition generating type countermeasure network
CN112148517A (en) * 2020-10-19 2020-12-29 东北电力大学 A fault diagnosis method for rotating equipment
EP3798917A1 (en) * 2019-09-24 2021-03-31 Naver Corporation Generative adversarial network (gan) for generating images

Patent Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2019056975A (en) * 2017-09-19 2019-04-11 株式会社Ｐｒｅｆｅｒｒｅｄ Ｎｅｔｗｏｒｋｓ Improved generative adversarial network achievement program, improved generative adversarial network achievement device, and learned model generation method
US20190197358A1 (en) * 2017-12-21 2019-06-27 International Business Machines Corporation Generative Adversarial Network Medical Image Generation for Training of a Classifier
CN108680807A (en) * 2018-05-17 2018-10-19 国网山东省电力公司青岛供电公司 The Diagnosis Method of Transformer Faults and system of network are fought based on condition production
CN108805418A (en) * 2018-05-22 2018-11-13 福州大学 A kind of traffic data fill method fighting network based on production
US20200202221A1 (en) * 2018-12-20 2020-06-25 Shandong University Of Science And Technology Fault detection method and system based on generative adversarial network and computer program
CN109918999A (en) * 2019-01-22 2019-06-21 西安交通大学 An intelligent diagnosis method for mechanical equipment faults based on generative model under small sample data
CN109813542A (en) * 2019-03-15 2019-05-28 中国计量大学 Fault Diagnosis Method of Air Handling Unit Based on Generative Adversarial Network
EP3798917A1 (en) * 2019-09-24 2021-03-31 Naver Corporation Generative adversarial network (gan) for generating images
CN110955651A (en) * 2019-11-28 2020-04-03 武汉科技大学 Motor fault data enhancement method based on deep convolution generation type countermeasure network
CN111445147A (en) * 2020-03-27 2020-07-24 中北大学 Generative confrontation network model evaluation method for mechanical fault diagnosis
CN111582348A (en) * 2020-04-29 2020-08-25 武汉轻工大学 Method, device, equipment and storage medium for training condition generating type countermeasure network
CN112148517A (en) * 2020-10-19 2020-12-29 东北电力大学 A fault diagnosis method for rotating equipment

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
刘云鹏;许自强;和家慧;王权;高树国;赵军;: "基于条件式Wasserstein生成对抗网络的电力变压器故障样本增强技术", 电网技术, no. 04 *
柴志豪;: "基于GAN的轴承故障诊断方法", 内燃机与配件, no. 14 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113900860A (en) * 2021-10-27 2022-01-07 重庆邮电大学 CGAN-based data recovery method for wireless sensor network fault node
CN115169506A (en) * 2022-09-06 2022-10-11 中铁第四勘察设计院集团有限公司 Method and system for rapidly diagnosing faults of power supply and transformation key equipment
CN116429406A (en) * 2023-06-14 2023-07-14 山东能源数智云科技有限公司 Construction method and device of fault diagnosis model of large-scale mechanical equipment
CN116429406B (en) * 2023-06-14 2023-09-26 山东能源数智云科技有限公司 Construction method and device of fault diagnosis model of large-scale mechanical equipment

Similar Documents

Publication Publication Date Title
CN110929763B (en) Mechanical fault diagnosis method of medium voltage vacuum circuit breaker based on multi-source data fusion
CN112286751B (en) An intelligent fault diagnosis system and method for high-end equipment based on edge-cloud collaboration
CN113505876A (en) High-voltage circuit breaker fault diagnosis method based on generation type countermeasure network
WO2023044978A1 (en) Adversarial-flow-model-based unsupervised fault diagnosis method for mechanical device
CN109298258A (en) Transformer fault diagnosis method and system combining RVM and DBN
CN113705615B (en) A multi-stage equipment fault diagnosis method and system for electric vehicle charging process based on neural network
CN113376516A (en) Medium-voltage vacuum circuit breaker operation fault self-diagnosis and early-warning method based on deep learning
CN107153735A (en) Motor driven systems PWM inverter method for diagnosing faults
CN110703077B (en) High-voltage circuit breaker fault diagnosis method based on HPSO-TSVM
CN114444620B (en) Indicator diagram fault diagnosis method based on generating type antagonistic neural network
CN114091549B (en) A device fault diagnosis method based on deep residual network
CN116842463A (en) A fault diagnosis method for electric vehicle charging pile equipment
CN115980560A (en) A CNN-GRU-based high-voltage circuit breaker mechanical fault diagnosis system, method, and equipment
CN116150635A (en) Unknown fault detection method for rolling bearings based on cross-domain correlation representation
Han et al. Using improved self-organizing map for partial discharge diagnosis of large turbogenerators
CN117056814A (en) Transformer voiceprint vibration fault diagnosis method
CN109901064B (en) ICA-LVQ-based high-voltage circuit breaker fault diagnosis method
CN110490218B (en) A self-learning method for rolling bearing faults based on two-stage DBN
CN112748331A (en) Circuit breaker mechanical fault identification method and device based on DS evidence fusion
CN112067053A (en) Multi-strategy joint fault diagnosis method for minority class identification
CN116032790B (en) Method, device and system for identifying, diagnosing and predicting massive data flow anomalies of dispatching automation system
CN118626898A (en) A cloud platform fault root cause analysis method based on abnormal correlation graph and graph neural network
CN117150399A (en) Novel fault identification method and device based on flow discrimination model
CN117810491A (en) Fuel cell system fault diagnosis method and system based on deep learning
CN110017989B (en) Method for diagnosing bearing fault of wind turbine

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20211015

RJ01 Rejection of invention patent application after publication