CN111537853A - Intelligent detection method for partial discharge of switch cabinet based on multi-source heterogeneous data analysis - Google Patents
Intelligent detection method for partial discharge of switch cabinet based on multi-source heterogeneous data analysis Download PDFInfo
- Publication number
- CN111537853A CN111537853A CN202010607236.7A CN202010607236A CN111537853A CN 111537853 A CN111537853 A CN 111537853A CN 202010607236 A CN202010607236 A CN 202010607236A CN 111537853 A CN111537853 A CN 111537853A
- Authority
- CN
- China
- Prior art keywords
- partial discharge
- network model
- switch cabinet
- data set
- dimensional data
- 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
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
- G01R31/1227—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Testing Relating To Insulation (AREA)
Abstract
The invention discloses a switch cabinet partial discharge intelligent detection method based on multi-source heterogeneous data analysis, which comprises the steps of obtaining historical operation data of a plurality of high-voltage switch cabinets which are already put into operation, and using the historical operation data as multi-source heterogeneous data related to partial discharge of the switch cabinets; extracting characteristic vectors from historical operating data and carrying out normalization processing to obtain a corresponding high-dimensional data set; carrying out dimensionality reduction on the total feature vector by utilizing deep learning to reduce the original high-dimensional data set into a low-dimensional data set; and training a partial discharge recognition network model by using a low-dimensional data set, and recognizing and detecting the partial discharge of the switch cabinet. According to the intelligent detection method for the partial discharge of the switch cabinet based on the multi-source heterogeneous data analysis, the intelligent identification of the partial discharge is realized by analyzing historical operation data, current, voltage, temperature, humidity, ultrasonic waves and transient voltage to earth data of the switch cabinet and utilizing a deep learning algorithm, the potential fault hazard of the switch cabinet is discovered in time, and the safety and stability of the switch cabinet are improved.
Description
Technical Field
The invention relates to the technical field of power equipment detection, in particular to a switch cabinet partial discharge intelligent detection method based on multi-source heterogeneous data analysis.
Background
The switch cabinet is one of the most important and complex devices in an electric power system, and in a power supply and distribution system of the electric power system, the switch cabinet is an important device mainly used for closing and opening an electric power circuit so as to transmit and switch an electric power load, and withdrawing a fault device and a line segment from the electric power system so as to ensure the safe operation of the system. As for the partial discharge detection method of the switch cabinet, there are mainly an ultrahigh frequency method, an ultrasonic detection method, a pulse current method, a temporary ground voltage method, and the like at present. These methods are to detect different signals generated by partial discharge, such as high frequency electromagnetic wave, pulse current, ultrasonic wave, etc., to determine whether the partial discharge occurs. However, the local discharge has various generation types and a generation position is not fixed, and meanwhile, the actual detection is influenced by various interferences due to a complicated electromagnetic environment on site, and the detection effect is difficult to meet the actual requirements. With the continuous development of smart power grids and the arrival of a big data era, information of various aspects of a large number of electrical equipment is readily available, and the data can reflect the running state information of the equipment from different sides, so that various identification characteristics are provided for the detection of partial discharge. However, such a large amount of data inevitably increases the difficulty of data processing, and this valuable resource cannot be fully utilized without proper algorithmic analysis processing.
Big data of the smart power grid have multisource heterogeneity, and various factors influencing partial discharge include local atmospheric temperature and humidity, namely meteorological data, and also include the size of the current of the busbar of the switch cabinet, namely the electric load at the moment, and are related to the insulation aging degree of the switch cabinet equipment, namely the time length of operation. And the storage departments and the data types of the data are different, so that the data have multi-source heterogeneity. How to handle the heterogeneous data of multisource will play an important role to improving cubical switchboard partial discharge detection effect.
In recent years, deep learning becomes one of important algorithms for big data processing, the feature quantity of data can be automatically extracted, the workload of manual extraction is reduced, the accuracy in the aspect of identification and classification is high, the advantages of a large amount of multi-source heterogeneous data can be fully utilized, and the local discharge detection effect of the switch cabinet is improved.
Disclosure of Invention
The invention aims to provide a switch cabinet partial discharge intelligent detection method based on multi-source heterogeneous data analysis.
In order to achieve the purpose, the invention provides the following scheme:
a switch cabinet partial discharge intelligent detection method based on multi-source heterogeneous data analysis comprises the following steps:
acquiring historical operating data of a plurality of high-voltage switch cabinets which are put into operation as multisource heterogeneous data related to partial discharge of the switch cabinets, wherein the historical operating data comprises current, voltage, temperature and humidity, ultrasonic waves, transient voltage-to-ground data and partial discharge occurrence conditions;
extracting feature vectors from the historical operating data and carrying out normalization processing to obtain high-dimensional feature vectors and obtain corresponding high-dimensional data sets;
thirdly, performing dimensionality reduction on the total feature vector by utilizing deep learning to reduce the original high-dimensional data set into a low-dimensional data set;
and step four, training a partial discharge recognition network model by using the low-dimensional data set, and recognizing and detecting the partial discharge of the switch cabinet by using the partial discharge recognition network model.
Optionally, in the first step, the current, voltage, temperature and humidity, the ultrasonic wave, and the transient ground voltage data are acquired by a current transformer, a voltage transformer, a temperature and humidity sensor, an ultrasonic sensor, and a transient ground voltage sensor, which are arranged in the switch cabinet, respectively.
Optionally, in the third step, the deep learning is used to perform the dimension reduction processing on the total feature vector, and the original high-dimensional data set is reduced to a low-dimensional data set, which specifically includes:
building a coding network model and a decoding network model, and training the coding network model and the decoding network model by using a high-dimensional data set;
and D, reducing the dimension of the high-dimensional data set in the step two into a low-dimensional data set by using the trained coding network model.
Optionally, the encoding network model and the decoding network model are trained at least 50 rounds using the high-dimensional dataset.
Optionally, in the fourth step, the partial discharge recognition network model is trained by using the low-dimensional data set, and the partial discharge recognition network model is used for recognizing and detecting the partial discharge of the switch cabinet, which specifically includes:
building a partial discharge recognition network model, and training the partial discharge recognition network model by using a low-dimensional data set generated by the trained coding network model;
and inputting real-time operation data of the switch cabinet to be detected into the trained partial discharge recognition network model to perform partial discharge recognition detection.
Optionally, the partial discharge recognition network model is trained for at least 20 rounds using a low-dimensional data set generated by the trained coding network model.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects: the intelligent detection method for the partial discharge of the switch cabinet based on the multi-source heterogeneous data analysis obtains multi-source heterogeneous data related to the partial discharge of the switch cabinet, wherein the multi-source heterogeneous data comprises historical operation data of current, voltage, temperature and humidity, ultrasonic waves, transient voltage-to-ground voltage data and partial discharge occurrence conditions of the switch cabinet; extracting characteristic vectors from various data and carrying out normalization processing to form a data set; reducing the dimension of the total feature vector by utilizing deep learning, and reducing the original high-dimensional data set into a low-dimensional data set; finally, training a partial discharge recognition network model by using a low-dimensional data set, and recognizing and detecting the partial discharge of the switch cabinet by using the partial discharge recognition network model; the method is different from the traditional switch cabinet partial discharge detection method, and according to various influence factors of the partial discharge of the switch cabinet, the corresponding multi-source heterogeneous data is obtained by utilizing the long-term operation data of the switch cabinet, so that the dimensionality of a large amount of high-dimensional data is reduced, the calculated amount is reduced, the detection difficulty is reduced, a road is paved for the later application of the method to edge calculation, and the various influence factors are fully considered, so that the partial discharge identification is more accurate, the insulation problem of the switch cabinet can be found more timely, and the operation stability of a power system transformer substation is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
Fig. 1 is a flowchart of an intelligent detection method for partial discharge of a switch cabinet based on multi-source heterogeneous data analysis according to an embodiment of the invention;
FIG. 2 illustrates an encoding network model and a decoding network model according to an embodiment of the present invention;
fig. 3 is a partial discharge identification network model according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention aims to provide a switch cabinet partial discharge intelligent detection method based on multi-source heterogeneous data analysis.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
Fig. 1 is a flowchart of an intelligent detection method for partial discharge of a switchgear based on multi-source heterogeneous data analysis according to an embodiment of the present invention, and as shown in fig. 1, the intelligent detection method for partial discharge of a switchgear based on multi-source heterogeneous data analysis according to an embodiment of the present invention includes the following steps:
acquiring historical operating data of a plurality of high-voltage switch cabinets which are put into operation as multisource heterogeneous data related to partial discharge of the switch cabinets, wherein the historical operating data comprises current, voltage, temperature and humidity, ultrasonic waves, transient voltage-to-ground data and partial discharge occurrence conditions;
extracting feature vectors from the historical operating data and carrying out normalization processing to obtain high-dimensional feature vectors and obtain corresponding high-dimensional data sets;
thirdly, performing dimensionality reduction on the total feature vector by utilizing deep learning to reduce the original high-dimensional data set into a low-dimensional data set;
and step four, training a partial discharge recognition network model by using the low-dimensional data set, and recognizing and detecting the partial discharge of the switch cabinet by using the partial discharge recognition network model.
In the first step, the current, voltage, temperature and humidity, ultrasonic waves and transient earth voltage data are acquired by a current transformer, a voltage transformer, a temperature and humidity sensor, an ultrasonic sensor and a transient earth voltage sensor which are arranged in the switch cabinet respectively.
In the second step, M pieces of insulation equipment are selected from the switch cabinet, and normalization processing is carried out on the commissioned time length data of each piece of equipment (the commissioning time length is divided by the statistical average service life of the equipment); carrying out linear normalization processing on the temperature and humidity data as shown in a formula (1); obtaining bus voltage by a voltage transformer and dividing the bus voltage by a rated voltage value of the switch cabinet, and obtaining a rated current value corresponding to the current value divided by the rated capacity by a current transformer; ultrasonic and discharge waveforms are obtained by ultrasonic and transient earth voltage sensors, characteristic parameters are extracted by utilizing the shape of discharge pulses, and the mean value v, the standard deviation sigma and the steepness K are respectively obtained by the formulas 1.2-1.5uInclination of Sk. And synthesizing the characteristic parameters to obtain a high-dimensional characteristic vector and obtain a corresponding data set. The high-dimensional characteristic vector comprises M-dimensional equipment operation duration, 2-dimensional temperature and humidity, 2-dimensional operation voltage and current, 4-dimensional ultrasonic waveform characteristic parameters and 4-dimensional transient ground voltage waveform characteristic parameters, and the total is M +12 dimensions.
In the formula (1), x is temperature and humidity data, min (x) and max (x) are the minimum value and the maximum value, and x' is the temperature and humidity after treatment. In the formula (2), the degree of skewness SKIs used to describe the degree of skewness of the distribution of the partial discharge pulse shape with respect to the normal distribution shape, and in the formula (5), the steepness KuThe degree of prominence of the distribution used to describe the shape of the partial discharge pulse compared to the shape of a normal distribution. Wherein v, sigma and pjRespectively, the mean, standard deviation, probability of the waveform amplitude within the phase window j. The mean and standard deviation are respectively formula (3) and formula (4), wherein [ + ]jIs the phase of the phase window, m is the number of phase windows, yjFor the amplitude of the jth phase window, Δ y is the difference between the adjacent phase window amplitudes.
In the third step, the deep learning is used for reducing the dimension of the total feature vector, and the original high-dimensional data set is reduced into a low-dimensional data set, which specifically comprises the following steps:
building a coding network model and a decoding network model, and training the coding network model and the decoding network model by using a high-dimensional data set;
and D, reducing the dimension of the high-dimensional data set in the step two into a low-dimensional data set by using the trained coding network model.
The encoding network model and the decoding network model are respectively used for reducing the original high-dimensional feature vector into a low-dimensional feature vector and restoring the low dimension into the original high-dimensional feature vector.
The coding neural network model comprises an input layer, a hidden layer and an output layer. As shown in fig. 2, the input layer is a single layer of neurons and includes M +12 input neurons, which refer to the high-dimensional feature vectors in step two. The hidden layer can be composed of a plurality of layers of neurons, the method is arranged into 2 layers, the first hidden layer comprises 20 neurons, and the second hidden layer comprises 10 neurons. The output layer is a single layer of neurons, including 5 output neurons, representing low-dimensional feature vectors. Initializing the weight w between each neuron to be 0, wherein the activation function of the hidden layer is tanh, and the output layer is a linear activation function. the tanh function is expressed by equation (6), and the learning rate lr is set to 0.01.
The decoding neural network model comprises an input layer, a hidden layer and an output layer. As shown in fig. 2, the input layer is a single layer of neurons and includes 5 input neurons, which refer to the low-dimensional feature vectors in step three. The hidden layer can be composed of a plurality of layers of neurons, the method is arranged into 2 layers, the first hidden layer comprises 10 neurons, and the second hidden layer comprises 20 neurons. The output layer is a single layer of neurons, including M +12 output neurons, representing high-dimensional feature vectors. Initializing the weight w between each neuron to be 0, wherein the activation function of the hidden layer is tanh, and the output layer is a sigmoid activation function. the tanh function is shown in formula (6), and the sigmoid function is shown in formula (7). The learning rate lr is set to 0.01.
Training the coding network and the decoding network in the step three by using the high-dimensional characteristic vector data set input in the step two, inputting the training data set into an input layer of a coding network model, and performing forward propagation, namely according to the output value x of a neuron in the front layeriWeight w between two layersijThe input value z of the layer is calculated by weightingjAs shown in equation 1.8, and the output value y of this layerjAnd (3) processing the input value of the layer by an activation function f, as shown in a formula (9), and calculating the output value of the output layer of the coding network finally by analogy, namely the low-dimensional data. Then inputting the output value of the output layer of the coding network into the input layer of the decoding network, and calculating the output value y of the output layer by the same forward propagation until the output layer of the decoding networklWill be associated with a high dimensional training data set ytThere is an error in the presence of the error,
namely, it is
Determining the partial derivative of the error in the network to each weight, i.e.
Backward propagation layer by layer, i.e. weight of each layer from wijIs modified into
Until the input layer of the coding network.
zj＝∑wijxi(8)
yj＝f(zj) (9)
The encoding network model and the decoding network model are trained at least 50 rounds using the high-dimensional dataset.
In the fourth step, the low-dimensional data set is utilized to train the partial discharge recognition network model, and the partial discharge recognition network model is utilized to recognize and detect the partial discharge of the switch cabinet, which specifically comprises the following steps:
building a partial discharge recognition network model, and training the partial discharge recognition network model by using a low-dimensional data set generated by the trained coding network model;
and inputting real-time operation data of the switch cabinet to be detected into the trained partial discharge recognition network model to perform partial discharge recognition detection.
The partial discharge recognition network model comprises an input layer, a hidden layer and an output layer. As shown in fig. 3, the input layer is a single layer of neurons and includes 5 input neurons, which refer to low-dimensional feature vectors. The hidden layer can be composed of a plurality of layers of neurons, the method is arranged into 3 layers, and each layer contains 10 neurons. The output layer is a single layer of neurons and comprises 1 output neuron, and the probability of partial discharge is shown. And initializing the weight w between each neuron to be 0, wherein the activation function of the hidden layer is relu, and the output layer is the sigmoid activation function. The relu function is shown in formula (10), and the sigmoid function is shown in formula (7). The learning rate lr is set to 0.01.
And inputting the low-dimensional data set into an input layer of the partial discharge recognition model as a training data set, and performing forward propagation and error backward propagation.
Wherein, the partial discharge recognition network model is trained for at least 20 rounds by utilizing a low-dimensional data set generated by the trained coding network model.
And deploying the coding network model and the partial discharge recognition network model trained in the third step and the fourth step in an actual switch cabinet, extracting high-dimensional characteristic vectors of real-time operation data of the switch cabinet to form a high-dimensional data set as shown in the second step, inputting the high-dimensional characteristic vectors and the high-dimensional data set into the coding network model to obtain low-dimensional characteristic vectors and a data set, inputting the low-dimensional characteristic vectors and the data set into the partial discharge recognition network model, and finally obtaining a recognition result.
The invention provides a switch cabinet partial discharge intelligent detection method based on multi-source heterogeneous data analysis, which comprises the steps of firstly obtaining historical operation data of a plurality of high-voltage switch cabinets which are already put into operation, processing various data and synthesizing characteristic parameters to obtain high-dimensional characteristic vectors, and obtaining corresponding data sets; building a coding and decoding network model, and training a coding and decoding network by using a high-dimensional feature vector data set; building a partial discharge recognition network model, and training the partial discharge recognition network by using a low-dimensional data set generated by the trained coding network; and deploying the trained coding network and the local discharge identification network in an actual switch cabinet, and completing the local discharge identification of the real-time data of the switch cabinet by using the model.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.
Claims (6)
1. A switch cabinet partial discharge intelligent detection method based on multi-source heterogeneous data analysis is characterized by comprising the following steps:
acquiring historical operating data of a plurality of high-voltage switch cabinets which are put into operation as multisource heterogeneous data related to partial discharge of the switch cabinets, wherein the historical operating data comprises current, voltage, temperature and humidity, ultrasonic waves, transient voltage-to-ground data and partial discharge occurrence conditions;
extracting feature vectors from the historical operating data and carrying out normalization processing to obtain high-dimensional feature vectors and obtain corresponding high-dimensional data sets;
thirdly, performing dimensionality reduction on the total feature vector by utilizing deep learning to reduce the original high-dimensional data set into a low-dimensional data set;
and step four, training a partial discharge recognition network model by using the low-dimensional data set, and recognizing and detecting the partial discharge of the switch cabinet by using the partial discharge recognition network model.
2. The intelligent detection method for the partial discharge of the switch cabinet based on the multi-source heterogeneous data analysis according to claim 1, wherein in the first step, the current, voltage, temperature and humidity, ultrasonic wave and transient voltage to ground voltage data are acquired by a current transformer, a voltage transformer, a temperature and humidity sensor, an ultrasonic sensor and a transient voltage to ground voltage sensor which are arranged in the switch cabinet respectively.
3. The intelligent detection method for partial discharge of switch cabinet based on multi-source heterogeneous data analysis according to claim 1, wherein in the third step, the deep learning is used for performing dimensionality reduction processing on the total feature vector, and an original high-dimensional data set is reduced to a low-dimensional data set, and the method specifically comprises:
building a coding network model and a decoding network model, and training the coding network model and the decoding network model by using a high-dimensional data set;
and D, reducing the dimension of the high-dimensional data set in the step two into a low-dimensional data set by using the trained coding network model.
4. The intelligent detection method for the partial discharge of the switch cabinet based on the multi-source heterogeneous data analysis according to claim 3, wherein the coding network model and the decoding network model are trained for at least 50 rounds by using a high-dimensional data set.
5. The intelligent detection method for the partial discharge of the switch cabinet based on the multi-source heterogeneous data analysis according to claim 1, wherein in the fourth step, a low-dimensional data set is used for training a partial discharge recognition network model, and the partial discharge recognition network model is used for recognizing and detecting the partial discharge of the switch cabinet, and specifically comprises:
building a partial discharge recognition network model, and training the partial discharge recognition network model by using a low-dimensional data set generated by the trained coding network model;
and inputting real-time operation data of the switch cabinet to be detected into the trained partial discharge recognition network model to perform partial discharge recognition detection.
6. The intelligent detection method for the partial discharge of the switch cabinet based on the multi-source heterogeneous data analysis according to claim 5, wherein the partial discharge recognition network model is trained for at least 20 rounds by using a low-dimensional data set generated by the trained coding network model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010607236.7A CN111537853A (en) | 2020-06-29 | 2020-06-29 | Intelligent detection method for partial discharge of switch cabinet based on multi-source heterogeneous data analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010607236.7A CN111537853A (en) | 2020-06-29 | 2020-06-29 | Intelligent detection method for partial discharge of switch cabinet based on multi-source heterogeneous data analysis |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111537853A true CN111537853A (en) | 2020-08-14 |
Family
ID=71976510
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010607236.7A Pending CN111537853A (en) | 2020-06-29 | 2020-06-29 | Intelligent detection method for partial discharge of switch cabinet based on multi-source heterogeneous data analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111537853A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112630562A (en) * | 2020-12-04 | 2021-04-09 | 国网四川省电力公司广安供电公司 | Switch cabinet fault identification method and device based on deep neural network |
| CN112748317A (en) * | 2021-03-23 | 2021-05-04 | 国网河南省电力公司电力科学研究院 | Switch cabinet partial discharge fault detection method and system based on multiple monitoring data |
| CN112881812A (en) * | 2021-01-11 | 2021-06-01 | 中国气象科学研究院 | Full-flash real-time positioning method and device based on machine learning coding |
| CN115355942A (en) * | 2022-08-16 | 2022-11-18 | 华珑(沈阳)智能电气有限公司 | Health management method suitable for KYN switch cabinet |
| CN115659248A (en) * | 2022-11-07 | 2023-01-31 | 中国长江三峡集团有限公司 | Power equipment defect identification method, device, equipment and storage medium |
| CN116026403A (en) * | 2022-11-09 | 2023-04-28 | 国能四川西部能源股份有限公司 | Switch cabinet fault early warning method and device, medium and electronic equipment |
| CN117312925A (en) * | 2023-11-24 | 2023-12-29 | 江苏征途技术股份有限公司 | Switch cabinet three-in-one partial discharge mode identification method based on improved AFT algorithm and BP neural network optimization |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102628917A (en) * | 2012-04-25 | 2012-08-08 | 广州供电局有限公司 | Partial discharge recognition method and system |
| CN104849633A (en) * | 2015-05-15 | 2015-08-19 | 国家电网公司 | Switchgear partial discharge mode recognition method |
| CN106503689A (en) * | 2016-11-14 | 2017-03-15 | 哈尔滨理工大学 | Neutral net local discharge signal mode identification method based on particle cluster algorithm |
| CN107831409A (en) * | 2017-09-22 | 2018-03-23 | 国网山东省电力公司电力科学研究院 | The method and method for detecting abnormality of superfrequency partial discharge detection TuPu method parameter extraction |
| CN109596955A (en) * | 2018-12-30 | 2019-04-09 | 国网北京市电力公司 | Shelf depreciation state determines method and device |
| CN110147806A (en) * | 2018-10-08 | 2019-08-20 | 腾讯科技（深圳）有限公司 | Training method, device and the storage medium of image description model |
| CN110244204A (en) * | 2019-06-27 | 2019-09-17 | 国网湖南省电力有限公司 | A kind of switchgear method for diagnosing faults, system and the medium of multiple characteristic values |
| CN110826934A (en) * | 2019-11-21 | 2020-02-21 | 广州供电局有限公司 | Method, device and system for evaluating health degree of medium-voltage switch cabinet |
| CN110988631A (en) * | 2019-12-16 | 2020-04-10 | 深圳江行联加智能科技有限公司 | Partial discharge positioning detection method and device, readable storage medium and electrical equipment |
-
2020
- 2020-06-29 CN CN202010607236.7A patent/CN111537853A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102628917A (en) * | 2012-04-25 | 2012-08-08 | 广州供电局有限公司 | Partial discharge recognition method and system |
| CN104849633A (en) * | 2015-05-15 | 2015-08-19 | 国家电网公司 | Switchgear partial discharge mode recognition method |
| CN106503689A (en) * | 2016-11-14 | 2017-03-15 | 哈尔滨理工大学 | Neutral net local discharge signal mode identification method based on particle cluster algorithm |
| CN107831409A (en) * | 2017-09-22 | 2018-03-23 | 国网山东省电力公司电力科学研究院 | The method and method for detecting abnormality of superfrequency partial discharge detection TuPu method parameter extraction |
| CN110147806A (en) * | 2018-10-08 | 2019-08-20 | 腾讯科技（深圳）有限公司 | Training method, device and the storage medium of image description model |
| CN109596955A (en) * | 2018-12-30 | 2019-04-09 | 国网北京市电力公司 | Shelf depreciation state determines method and device |
| CN110244204A (en) * | 2019-06-27 | 2019-09-17 | 国网湖南省电力有限公司 | A kind of switchgear method for diagnosing faults, system and the medium of multiple characteristic values |
| CN110826934A (en) * | 2019-11-21 | 2020-02-21 | 广州供电局有限公司 | Method, device and system for evaluating health degree of medium-voltage switch cabinet |
| CN110988631A (en) * | 2019-12-16 | 2020-04-10 | 深圳江行联加智能科技有限公司 | Partial discharge positioning detection method and device, readable storage medium and electrical equipment |
Non-Patent Citations (4)
| Title |
|---|
| 刘玉良 等: "《深度学习》", 31 January 2020 * |
| 曾文 等: "基于深度学习技术的海量科技文献向量表示方法研究", 《2017年北京科学技术情报学会年会--"科技情报发展助力科技创新中心建设"论坛》 * |
| 桑塔努•帕塔纳雅克 等: "《TensorFlow深度学习：数学原理与Python实战进阶》", 30 April 2020 * |
| 贵州电网有限责任公司: "《GIS变电站运维中绝缘关键技术》", 31 December 2017 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112630562A (en) * | 2020-12-04 | 2021-04-09 | 国网四川省电力公司广安供电公司 | Switch cabinet fault identification method and device based on deep neural network |
| CN112881812A (en) * | 2021-01-11 | 2021-06-01 | 中国气象科学研究院 | Full-flash real-time positioning method and device based on machine learning coding |
| CN112881812B (en) * | 2021-01-11 | 2022-02-22 | 中国气象科学研究院 | Full-flash real-time positioning method and device based on machine learning coding |
| CN112748317A (en) * | 2021-03-23 | 2021-05-04 | 国网河南省电力公司电力科学研究院 | Switch cabinet partial discharge fault detection method and system based on multiple monitoring data |
| CN112748317B (en) * | 2021-03-23 | 2022-03-25 | 国网河南省电力公司电力科学研究院 | Switch cabinet partial discharge fault detection method and system based on multiple monitoring data |
| CN115355942A (en) * | 2022-08-16 | 2022-11-18 | 华珑(沈阳)智能电气有限公司 | Health management method suitable for KYN switch cabinet |
| CN115659248A (en) * | 2022-11-07 | 2023-01-31 | 中国长江三峡集团有限公司 | Power equipment defect identification method, device, equipment and storage medium |
| CN116026403A (en) * | 2022-11-09 | 2023-04-28 | 国能四川西部能源股份有限公司 | Switch cabinet fault early warning method and device, medium and electronic equipment |
| CN117312925A (en) * | 2023-11-24 | 2023-12-29 | 江苏征途技术股份有限公司 | Switch cabinet three-in-one partial discharge mode identification method based on improved AFT algorithm and BP neural network optimization |
| CN117312925B (en) * | 2023-11-24 | 2024-03-01 | 浙江大学 | Switch cabinet three-in-one partial discharge mode identification method based on improved AFT algorithm and BP neural network optimization |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111537853A (en) | Intelligent detection method for partial discharge of switch cabinet based on multi-source heterogeneous data analysis | |
| Mishra et al. | A novel fault classification scheme for series capacitor compensated transmission line based on bagged tree ensemble classifier | |
| Zhang et al. | Fault diagnosis of power grid based on variational mode decomposition and convolutional neural network | |
| Sarlak et al. | High impedance fault detection using combination of multi-layer perceptron neural networks based on multi-resolution morphological gradient features of current waveform | |
| Luo et al. | Stacked auto-encoder based fault location in VSC-HVDC | |
| Xiao et al. | High-impedance fault detection method based on one-dimensional variational prototyping-encoder for distribution networks | |
| CN103323755A (en) | Method and system for recognition of GIS ultrahigh frequency partial discharge signal | |
| Ferreira et al. | Probabilistic transmission line fault diagnosis using autonomous neural models | |
| CN115564331A (en) | Power distribution network typical fault risk sensing method, power distribution network, equipment and storage medium | |
| CN110414412A (en) | The multiple disturbance precise recognition method of the Wide Area Power based on big data analysis and device | |
| CN110609477A (en) | Electric power system transient stability discrimination system and method based on deep learning | |
| CN105203869A (en) | Microgrid island detection method based on extreme learning machine | |
| CN114325256A (en) | Power equipment partial discharge identification method, system, equipment and storage medium | |
| Panigrahi et al. | A power quality detection and classification algorithm based on FDST and hyper-parameter tuned light-GBM using memetic firefly algorithm | |
| CN110068748A (en) | 10kV single-core cable early stage state identification method based on depth convolutional neural networks | |
| Kanwal et al. | Artificial intelligence based faults identification, classification, and localization techniques in transmission lines-a review | |
| Li et al. | Forecasting of wind capacity ramp events using typical event clustering identification | |
| Jiang et al. | A new method for recognition and classification of power quality disturbances based on IAST and RF | |
| CN115600754A (en) | Method, system, equipment and storage medium for predicting potential safety hazard of user power utilization | |
| CN113902581B (en) | Power consumption anomaly detection method based on depth self-encoder Gaussian mixture model | |
| Du et al. | A hierarchical power system transient stability assessment method considering sample imbalance | |
| Venkata et al. | Data mining and svm based fault diagnostic analysis in modern power system using time and frequency series parameters calculated from full-cycle moving window | |
| Luo et al. | Stacked auto-encoder-based transients recognition in VSC-HVDC | |
| Wei et al. | Deep Belief network based faulty feeder detection of single-phase ground fault | |
| Muzzammel et al. | Wavelet multiscale principal components and traveling waves enabled machine learning technique for protection of MT-HVDC systems |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20201120 Address after: 274000 north side of Zhonghua East Road, Heze, Shandong Applicant after: Heze Power Supply Company of State Grid Shandong Electric Power Company Applicant after: STATE GRID CORPORATION OF CHINA Address before: 274000 north side of Zhonghua East Road, Heze, Shandong Applicant before: Heze Power Supply Company of State Grid Shandong Electric Power Company |
|
| TA01 | Transfer of patent application right | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200814 |
|
| RJ01 | Rejection of invention patent application after publication |