CN110927535A - Power equipment partial discharge severity evaluation method based on extreme learning machine - Google Patents
Power equipment partial discharge severity evaluation method based on extreme learning machine Download PDFInfo
- Publication number
- CN110927535A CN110927535A CN201911097771.6A CN201911097771A CN110927535A CN 110927535 A CN110927535 A CN 110927535A CN 201911097771 A CN201911097771 A CN 201911097771A CN 110927535 A CN110927535 A CN 110927535A
- Authority
- CN
- China
- Prior art keywords
- partial discharge
- image
- learning machine
- extreme learning
- power equipment
- 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
-
- 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
- G01R31/1218—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 using optical methods; using charged particle, e.g. electron, beams or X-rays
-
- 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
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Testing Relating To Insulation (AREA)
Abstract
The invention relates to a method for evaluating the severity of partial discharge of electrical equipment based on an extreme learning machine, which is technically characterized by comprising the following steps of: dividing the severity of partial discharge of the power equipment; performing dimensionality reduction processing on the partial discharge detection image data of the power equipment; constructing a neural network containing a plurality of hidden layers through a stack type automatic encoder, extracting deeper features of target data, and mining essential information of the data; inputting the target data deep-level features and operation related data information extracted by the stack type automatic encoder into the extreme learning machine to obtain output layer result information based on the extreme learning machine; and comparing the consistency of the actual maintenance result and the evaluation result, and training and learning the evaluation model. The invention has reasonable design, realizes the function of evaluating the partial discharge severity of the power equipment based on the extreme learning machine model, and meets the requirement of carrying out management and maintenance work more accurately.
Description
Technical Field
The invention belongs to the technical field of high-voltage insulation state evaluation, and particularly relates to a method for evaluating partial discharge severity of electrical equipment based on an extreme learning machine.
Background
The current partial discharge research of the power equipment mainly comprises the steps of building an environment in a laboratory, simulating the change of the severity of defects by using a step pressurization method, and analyzing the characteristics of collected partial discharge signals. Selars et al simulate discharge such as particulate discharge, point discharge, insulator surface discharge, solid insulation internal discharge and the like in a GIS under an experimental environment, and analyze detection signal characteristics of the discharge at different development stages by using a UHF detection method. The homodyne and the like use a pulse current method, a ultrahigh frequency method, an ultrasonic method and an optical method to research the surface fixed metal particles along the surface of a GIS insulator and the fault partial discharge of a high-voltage electrode of the insulator in the experimental environment, divide the severity of different partial discharges and analyze the data characteristics.
With the deep progress of the state detection and state evaluation work of the power grid company, ultrahigh frequency, ultrasonic wave and high frequency detection technologies are widely applied. Because the partial discharge detection data is lack of a uniform format, instrument manufacturers often customize the type and format of data storage according to the characteristics of the instruments, most manufacturers only store picture formats and do not support data retrieval; the defect judgment based on the three detection technologies is mainly based on a typical map comparison mode. Due to the fact that the site condition of the transformer substation is complex, the position of a partial discharge source, the running voltage of a GIS, the load current, the equipment operation time, the equipment structure and other information can affect the severity of partial discharge, and due to the fact that no good method exists for evaluating the severity of detected partial discharge signals, management and maintenance work cannot be conducted more accurately.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provide the method for evaluating the partial discharge severity of the electrical equipment based on the extreme learning machine, which is reasonable in design, accurate and reliable.
The technical problem to be solved by the invention is realized by adopting the following technical scheme:
a method for evaluating the severity of partial discharge of electric power equipment based on an extreme learning machine comprises the following steps:
step 1, dividing the severity of partial discharge of the power equipment;
step 2, performing dimensionality reduction processing on the partial discharge detection image data of the power equipment;
step 3, constructing a neural network containing a plurality of layers of hidden layers through a stack type automatic encoder, extracting deeper features of target data, and mining essential information of the data;
step 4, inputting the target data deep-level features and the operation related data information extracted by the stack type automatic encoder into the extreme learning machine to obtain output layer result information based on the extreme learning machine;
and 5, performing evaluation model training and learning according to the consistency comparison between the actual overhaul result and the evaluation result.
Further, step 1 is to classify the severity of the partial discharge of the power equipment as follows according to whether the equipment needs to be overhauled and the time requirement:
in general: no equipment fault occurs within one year after the partial discharge signal is detected or no obvious discharge trace is found after the partial discharge signal is disintegrated;
severe: equipment failure occurs within three months to one year after the partial discharge signal is detected or obvious discharge traces are found after the equipment failure or the disintegration is detected;
emergency: and after the equipment failure occurs within three months after the partial discharge signal is detected or the equipment is disassembled, obvious discharge traces are found.
Further, the specific implementation method of step 2 includes the following steps:
step 2-1, segmenting and extracting the image format data according to the image characteristics, removing irrelevant information such as channels, marks and the like, and finally obtaining a color image only containing pulse voltage, wherein the color image consists of color components of R, G, B three channels;
2-2, performing graying processing on the color image by adopting a component method to obtain waveform image gray maps of three channels;
step 2-3, carrying out global binarization operation on the waveform image gray level image, removing interference channels and other irrelevant information in the image, and only extracting and reserving to obtain green channels to obtain a binary image;
step 2-4, performing image enhancement processing to obtain a time domain waveform image;
step 2-5: compressing the time domain waveform image into an image of uniform size;
step 2-6: and carrying out sparse noise reduction on the image obtained by preprocessing.
Further, the image obtained in the step 2-5 is 96 × 96.
Further, the step 3 is specifically implemented as follows:
first, a label-free data set is given
Each of which training data x(i)Obtaining the feature expression y of the hidden layer through the operation of the encoder(i)：
y(i)＝fθ(x(i))＝s(Wx(i)+b)
Where θ ═ (W, b) is a network parameter, W is a weight matrix, b is a bias vector, s (x) is an activation function, and the activation function is a sigmoid function;
then the characteristic expression of the hidden layer is operated by a decoder to obtain a reconstructed vector z(i)：
z(i)＝gθ′(y(i))＝s(W′y(i)+b′)
Where θ ═ W ', b ', W ' is a weight matrix, and W ═ W is takenTThen the model actually needs to minimize the cost function as:
adding a regularization to the cost function to obtain a final cost function as follows:
further, the extreme learning machine is composed of an input layer, a single hidden layer and an output layer, and the connection weight between the input layer and the hidden layer and the threshold value of a hidden layer neuron are randomly generated.
Further, the operation related data information in step 4 includes equipment account information, equipment nameplate information, technical parameters of the detection instrument, equipment operation data and weather information.
Further, the specific implementation method of step 5 is as follows: performing equipment management maintenance according to the definition of the partial discharge severity of the power equipment corresponding to the result information, and after the equipment management maintenance, performing consistency comparison on an on-site actual result and an evaluation result; and if the results are consistent, the evaluation results are accurate, if the results are inconsistent, the results are fed back to the evaluation model based on the extreme learning machine for training and learning, and the data model is corrected until the evaluation results are consistent with the actual results.
The invention has the advantages and positive effects that:
according to the method, deeper features of the target data are extracted by utilizing the fitting data of the self-encoder, the data information related to operation is input to the extreme learning machine model to evaluate the partial discharge severity of the power equipment with different data formats, and the training and learning promotion of the evaluation model can be carried out according to the consistency of the later actual overhaul result and the evaluation result, so that the requirement of carrying out management and maintenance work more accurately is met.
Drawings
FIG. 1 is a flow chart of the present invention for evaluating the severity of partial discharge of an electric power device based on an extreme learning machine;
FIG. 2 is a flow of image data pre-processing;
FIG. 3 is a waveform image segmentation;
FIG. 4 is a waveform image gray scale map (R channel);
FIG. 5 is a waveform image gray scale map (G channel);
FIG. 6 is a waveform image gray scale map (B channel);
FIG. 7 is a binary image of a waveform image;
FIG. 8 is an expanded view of a first waveform image;
FIG. 9 is an expanded view of a second waveform image;
fig. 10 is a diagram of a self-encoder structure.
Detailed Description
The following describes the embodiments of the present invention in detail with reference to the accompanying drawings.
The invention provides a method for evaluating partial discharge severity of electric equipment based on an Extreme Learning Machine (ELM), which comprises the following steps as shown in figure 1:
step 1, dividing the severity of partial discharge of the power equipment.
Because the defect judgment of the traditional detection technology based on the ultrahigh frequency, the ultrasonic wave and the high frequency is mainly based on a typical map comparison mode, no good method exists for evaluating the severity of the detected partial discharge signal, and the management and maintenance work cannot be carried out more accurately.
Therefore, the invention divides the severity of the partial discharge of the power equipment according to whether the equipment needs to be overhauled and the time requirement. The partial discharge severity of the power equipment is defined as "general", "severe", and "critical".
"general": no equipment failure occurs within one year after the partial discharge signal is detected or no obvious discharge trace is found after the partial discharge signal is disintegrated.
"severe": and after the partial discharge signal is detected, equipment failure occurs within three months to one year or obvious discharge traces are found after the equipment failure or the disassembly.
"critical": and after the equipment failure occurs within three months after the partial discharge signal is detected or the equipment is disassembled, obvious discharge traces are found.
And 2, performing dimensionality reduction on the partial discharge detection image data of the power equipment.
At present, instruments used in substation field partial discharge detection are various, existing stored partial discharge data mostly exist in the form of images, and a characteristic value-based partial discharge pattern recognition method cannot be widely applied to unstructured data. Therefore, in the step, the image processing technology is firstly adopted to preprocess the partial discharge image information data, and then the depth sparse noise reduction image information is obtained. The specific processing process of the image preprocessing comprises the following steps: image segmentation extraction, image graying, image binarization, image enhancement and image compression, as shown in fig. 2.
The specific treatment process of the step is as follows:
step 2-1: the image format data is divided and extracted according to the image characteristics, irrelevant information such as channels, marks and the like is removed, and finally an image only containing pulse voltage is obtained, as shown in fig. 3.
Step 2-2: the color image obtained in the step 2-1 is composed of color components of R, G, B channels, and the color image is subjected to gray scale processing by a component method, wherein the processing method comprises the following steps:
f1(i,j)＝R(i,j)；f2(i,j)＝G(i,j)；f3(i,j)＝B(i,j) (1)
wherein f isk(i, j) (k ═ 1, 2, 3) is the grayscale value of the converted grayscale image at (i, j). The waveform image gray-scale maps of the three channels as shown in fig. 4, 5 and 6 are obtained.
Step 2-3: and (3) carrying out global binarization operation on the waveform image gray level image obtained in the step (2-2), removing an interference channel and other irrelevant information in the image, and only extracting and reserving to obtain a green channel, wherein the processing method comprises the following steps:
the binary image shown in fig. 7 is obtained after the processing in this step.
Step 2-4: and enhancing the image to obtain a time domain waveform image.
Because the sample data obtained by detection is less and the sample data sets of different types are unbalanced, the performance of a machine learning algorithm is easy to be worsened, and overfitting occurs in the process of network training and learning. In order to reduce overfitting and enhance the network generalization capability, the patent adopts a Data augmentation (Data augmentation) method, and the Data set is artificially enlarged by cutting the image in a sliding way. And selecting different sliding step lengths aiming at the defect images with different sample sizes, and expanding according to a time domain waveform image to obtain new samples with different quantities. For example, the present patent expands a part of the interference type image into two images as a new sample, and the processed images are shown in fig. 8 and 9.
Step 2-5: image compression: the time domain waveform images are uniformly compressed into 96 × 96 images.
Through image enhancement, sample expansion is carried out on the image, and finally 8568 time domain waveform images are obtained in total. Because the configuration of the identification experiment platform is limited, the time domain waveform image obtained by preprocessing is compressed into an image with the size of 96 multiplied by 96 in a unified way.
Step 2-6: and carrying out sparse noise reduction processing on the image data obtained by preprocessing so as to facilitate data unification.
And 3, constructing a neural network containing multiple hidden layers through a stacked Auto-Encoder (SAE), mapping the characteristic data, unsupervised acquiring deep essential information of the data, extracting deeper characteristics of the target data and mining the essential information of the data.
Hinton toIn 2006, a data denoising deep learning idea is proposed, and deep intrinsic information of data is obtained by constructing a neural network containing a plurality of hidden layers and characteristic mapping data. A stacked Auto-Encoder (SAE) is used as an important component of a deep learning structure, and can extract deeper features of target data and mine essential information of the data. The autoencoder is an unsupervised learning algorithm, and its structure is shown in FIG. 10, by making the target value equal to the input value, i.e., y(i)＝x(i)And is used for learning nonlinear codes capable of recovering the nonlinear codes.
If given a label-free data set
Each of which training data x(i)By the operation of the encoder, the characteristic expression y of the hidden layer can be obtained(i)，y(i)＝fθ(x(i))＝s(Wx(i)+ b), where θ ═ (W, b) is the network parameter, W is the weight matrix, b is the offset vector, s (x) is the activation function, sigmoid function is selected, s (x) is 1/(1+ e)-x). Then the characteristic expression of the hidden layer is operated by a decoder to obtain a reconstructed vector z(i)，z(i)＝zθ′(y(i))＝s(W′y(i)+ b '), where θ' ═ (W ', b'), and W 'is a weight matrix, and W' ═ W is takenT. The model actually needs to minimize the cost function (this patent uses cross entropy cost function, better than squared error cost function):
in order to prevent the over-fitting situation from occurring, a regularization needs to be added to the cost function, and the obtained cost function is:
and 4, step 4: and inputting the target data deep-level features and the operation related data information extracted by the stack type automatic encoder into the extreme learning machine ELM to obtain output layer result information based on the extreme learning machine ELM.
In this step, the relevant data of the equipment operation mainly includes relevant operation information of the equipment with defects and faults during the detection of the defects and faults, including equipment account information, equipment name plate information, technical parameters of detection instruments, equipment operation data, meteorological information and the like. These operation-related data information are important parameters for evaluating the severity of partial discharge of the power equipment.
In the step, an Extreme Learning Machine (ELM) is used as a classifier, and because the connection weight between the input layer and the hidden layer and the threshold of the hidden layer neuron are randomly generated, iteration and adjustment are not needed in the training process, and the method has the advantages of high Learning rate, good generalization performance and the like.
The ELM consists of an input layer, a single hidden layer and an output layer. The connection weight between the input layer and the hidden layer and the threshold value of the hidden layer neuron are randomly generated, and iteration and adjustment are not needed in the training process. N sample data X ═ X is set1，x2，…，xn]T，xi∈RQAnd Q is the dimension of the input feature vector. The output matrix is T ═ T1，t2，…,tn]T，ti∈RSAnd S is the dimension of the output vector. Setting hidden layer activation function as g (x), randomly generating weight and threshold (w) from input layer to hidden layeri，bi)，wi，bi∈(-1，1)，
The number of neurons in the hidden layer is implied. The hidden layer output is then:
it is desirable to find β that satisfies H β -T minimum, where β is the weight of the connection of the hidden layer to the output layer, the infinitely differentiable activation function g (x) in any interval, R → R, at the randomly assigned inputLayer-to-hidden layer weight matrix wiAnd a threshold value biIn the case of (2), when the number of neurons in the hidden layer is the same as the number of training sample data, H is invertible in a broad sense, and β exists so that | | H β -T | | ═ 0, so that:
β＝H-1T (6)
and 5: and performing equipment management maintenance according to the definition of the partial discharge severity of the power equipment corresponding to the result information, and performing evaluation model training, learning and promotion according to the consistency detection of the later-stage actual overhaul result and the evaluation result.
In this step, after the device management and maintenance, the on-site actual result is compared with the evaluation result in consistency. If the results are consistent, the evaluation result is accurate; and if the results are inconsistent, the data are fed back to the evaluation model based on the extreme learning machine for training and learning, and the data model is corrected until the evaluation result is consistent with the actual result, so that the purpose of intelligently improving the learning of the model is achieved.
Nothing in this specification is said to apply to the prior art.
It should be emphasized that the embodiments described herein are illustrative rather than restrictive, and thus the present invention is not limited to the embodiments described in the detailed description, but also includes other embodiments that can be derived from the technical solutions of the present invention by those skilled in the art.
Claims (8)
1. A method for evaluating the severity of partial discharge of electric power equipment based on an extreme learning machine is characterized by comprising the following steps:
step 1, dividing the severity of partial discharge of the power equipment;
step 2, performing dimensionality reduction processing on the partial discharge detection image data of the power equipment;
step 3, constructing a neural network containing a plurality of layers of hidden layers through a stack type automatic encoder, extracting deeper features of target data, and mining essential information of the data;
step 4, inputting the target data deep-level features and the operation related data information extracted by the stack type automatic encoder into the extreme learning machine to obtain output layer result information based on the extreme learning machine;
and 5, performing evaluation model training and learning according to the consistency comparison between the actual overhaul result and the evaluation result.
2. The extreme learning machine-based power equipment partial discharge severity assessment method according to claim 1, characterized in that: step 1 is to divide the severity of the partial discharge of the power equipment into the following parts according to whether the equipment needs to be overhauled and the time requirement:
in general: no equipment fault occurs within one year after the partial discharge signal is detected or no obvious discharge trace is found after the partial discharge signal is disintegrated;
severe: equipment failure occurs within three months to one year after the partial discharge signal is detected or obvious discharge traces are found after the equipment failure or the disintegration is detected;
emergency: and after the equipment failure occurs within three months after the partial discharge signal is detected or the equipment is disassembled, obvious discharge traces are found.
3. The extreme learning machine-based power equipment partial discharge severity assessment method according to claim 1, characterized in that: the specific implementation method of the step 2 comprises the following steps:
step 2-1, segmenting and extracting the image format data according to the image characteristics, removing irrelevant information such as channels, marks and the like, and finally obtaining a color image only containing pulse voltage, wherein the color image consists of color components of R, G, B three channels;
2-2, performing graying processing on the color image by adopting a component method to obtain waveform image gray maps of three channels;
step 2-3, carrying out global binarization operation on the waveform image gray level image, removing interference channels and other irrelevant information in the image, and only extracting and reserving to obtain green channels to obtain a binary image;
step 2-4, performing image enhancement processing to obtain a time domain waveform image;
step 2-5: compressing the time domain waveform image into an image of uniform size;
step 2-6: and carrying out sparse noise reduction on the image obtained by preprocessing.
4. The extreme learning machine-based power equipment partial discharge severity assessment method according to claim 3, characterized in that: the image obtained in the step 2-5 is 96 multiplied by 96.
5. The extreme learning machine-based power equipment partial discharge severity assessment method according to claim 1, characterized in that: the specific implementation method of the step 3 is as follows:
first, a label-free data set is given
Each of which training data x(i)Obtaining the feature expression y of the hidden layer through the operation of the encoder(i)：
y(i)＝fθ(x(i))＝s(Wx(i)+b)
Where θ ═ (W, b) is a network parameter, W is a weight matrix, b is a bias vector, s (x) is an activation function, and the activation function is a sigmoid function;
then the characteristic expression of the hidden layer is operated by a decoder to obtain a reconstructed vector z(i)：
z(i)＝gθ′(y(i))＝s(W′y(i)+b′)
Where θ ═ W ', b ', W ' is a weight matrix, and W ═ W is takenTThen the model actually needs to minimize the cost function as:
adding a regularization to the cost function to obtain a final cost function as follows:
6. the extreme learning machine-based power equipment partial discharge severity assessment method according to claim 1, characterized in that: the extreme learning machine consists of an input layer, a single hidden layer and an output layer, wherein the connection weight between the input layer and the hidden layer and the threshold of a neuron of the hidden layer are randomly generated.
7. The extreme learning machine-based power equipment partial discharge severity assessment method according to claim 1, characterized in that: the operation related data information in the step 4 comprises equipment account information, equipment nameplate information, technical parameters of a detection instrument, equipment operation data and meteorological information.
8. The extreme learning machine-based power equipment partial discharge severity assessment method according to claim 1, characterized in that: the specific implementation method of the step 5 is as follows: performing equipment management maintenance according to the definition of the partial discharge severity of the power equipment corresponding to the result information, and after the equipment management maintenance, performing consistency comparison on an on-site actual result and an evaluation result; and if the results are consistent, the evaluation results are accurate, if the results are inconsistent, the results are fed back to the evaluation model based on the extreme learning machine for training and learning, and the data model is corrected until the evaluation results are consistent with the actual results.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911097771.6A CN110927535A (en) | 2019-11-12 | 2019-11-12 | Power equipment partial discharge severity evaluation method based on extreme learning machine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911097771.6A CN110927535A (en) | 2019-11-12 | 2019-11-12 | Power equipment partial discharge severity evaluation method based on extreme learning machine |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110927535A true CN110927535A (en) | 2020-03-27 |
Family
ID=69852686
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911097771.6A Pending CN110927535A (en) | 2019-11-12 | 2019-11-12 | Power equipment partial discharge severity evaluation method based on extreme learning machine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110927535A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111579939A (en) * | 2020-04-23 | 2020-08-25 | 天津大学 | Method for detecting partial discharge phenomenon of high-voltage power cable based on deep learning |
| CN111613340A (en) * | 2020-05-21 | 2020-09-01 | 安徽理工大学 | Miner health assessment method and system |
| CN111626153A (en) * | 2020-05-13 | 2020-09-04 | 电子科技大学 | Integrated learning-based partial discharge fault state identification method |
| CN111800301A (en) * | 2020-08-20 | 2020-10-20 | 浙江璟锐科技有限公司 | Network security evaluation method and system in machine type communication |
| CN112379235A (en) * | 2020-10-10 | 2021-02-19 | 广西电网有限责任公司电力科学研究院 | GIS insulation fault testability evaluation method |
| CN112686182A (en) * | 2021-01-04 | 2021-04-20 | 华北电力大学（保定） | Partial discharge mode identification method and terminal equipment |
| CN113033632A (en) * | 2021-03-10 | 2021-06-25 | 西北工业大学 | High-altitude platform fault diagnosis method based on wavelet analysis and multi-layer overrun learning machine |
| CN113342474A (en) * | 2021-06-29 | 2021-09-03 | 中国农业银行股份有限公司 | Method, device and storage medium for forecasting customer flow and training model |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20130129524A (en) * | 2012-05-21 | 2013-11-29 | 김현숙 | Fomentation medium having herb and bath tube system |
| CN105021957A (en) * | 2015-08-03 | 2015-11-04 | 西南石油大学 | Power cable accessory fault identification method and system |
| CN105548832A (en) * | 2015-12-10 | 2016-05-04 | 国网四川省电力公司电力科学研究院 | High-voltage power cable fault recognition method |
| CN108730776A (en) * | 2018-06-22 | 2018-11-02 | 大连理工大学 | A kind of Oil Pipeline Leakage Detection based on the fusion of extreme learning machine information |
| CN109001602A (en) * | 2018-08-20 | 2018-12-14 | 广东工业大学 | Shelf depreciation Severity method based on extreme learning machine algorithm |
-
2019
- 2019-11-12 CN CN201911097771.6A patent/CN110927535A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20130129524A (en) * | 2012-05-21 | 2013-11-29 | 김현숙 | Fomentation medium having herb and bath tube system |
| CN105021957A (en) * | 2015-08-03 | 2015-11-04 | 西南石油大学 | Power cable accessory fault identification method and system |
| CN105548832A (en) * | 2015-12-10 | 2016-05-04 | 国网四川省电力公司电力科学研究院 | High-voltage power cable fault recognition method |
| CN108730776A (en) * | 2018-06-22 | 2018-11-02 | 大连理工大学 | A kind of Oil Pipeline Leakage Detection based on the fusion of extreme learning machine information |
| CN109001602A (en) * | 2018-08-20 | 2018-12-14 | 广东工业大学 | Shelf depreciation Severity method based on extreme learning machine algorithm |
Non-Patent Citations (1)
| Title |
|---|
| 朱旭亮 等: "基于图像处理及降噪的局部放电图谱智能识别方法", 《电力大数据》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111579939A (en) * | 2020-04-23 | 2020-08-25 | 天津大学 | Method for detecting partial discharge phenomenon of high-voltage power cable based on deep learning |
| CN111626153A (en) * | 2020-05-13 | 2020-09-04 | 电子科技大学 | Integrated learning-based partial discharge fault state identification method |
| CN111626153B (en) * | 2020-05-13 | 2022-10-18 | 电子科技大学 | Integrated learning-based partial discharge fault state identification method |
| CN111613340A (en) * | 2020-05-21 | 2020-09-01 | 安徽理工大学 | Miner health assessment method and system |
| CN111800301A (en) * | 2020-08-20 | 2020-10-20 | 浙江璟锐科技有限公司 | Network security evaluation method and system in machine type communication |
| CN112379235A (en) * | 2020-10-10 | 2021-02-19 | 广西电网有限责任公司电力科学研究院 | GIS insulation fault testability evaluation method |
| CN112379235B (en) * | 2020-10-10 | 2022-06-24 | 广西电网有限责任公司电力科学研究院 | GIS insulation fault testability evaluation method |
| CN112686182A (en) * | 2021-01-04 | 2021-04-20 | 华北电力大学（保定） | Partial discharge mode identification method and terminal equipment |
| CN112686182B (en) * | 2021-01-04 | 2023-12-26 | 华北电力大学（保定） | Partial discharge mode identification method and terminal equipment |
| CN113033632A (en) * | 2021-03-10 | 2021-06-25 | 西北工业大学 | High-altitude platform fault diagnosis method based on wavelet analysis and multi-layer overrun learning machine |
| CN113342474A (en) * | 2021-06-29 | 2021-09-03 | 中国农业银行股份有限公司 | Method, device and storage medium for forecasting customer flow and training model |
| CN113342474B (en) * | 2021-06-29 | 2024-04-30 | 中国农业银行股份有限公司 | Method, equipment and storage medium for predicting customer flow and training model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110927535A (en) | Power equipment partial discharge severity evaluation method based on extreme learning machine | |
| Peng et al. | Random forest based optimal feature selection for partial discharge pattern recognition in HV cables | |
| CN108573225B (en) | Partial discharge signal pattern recognition method and system | |
| CN111353395B (en) | Face-changing video detection method based on long-term and short-term memory network | |
| CN109145961B (en) | Pattern recognition method and system for unstructured partial discharge data | |
| CN106778595B (en) | Method for detecting abnormal behaviors in crowd based on Gaussian mixture model | |
| Cord et al. | Automatic road defect detection by textural pattern recognition based on AdaBoost | |
| CN114841972B (en) | Transmission line defect identification method based on saliency map and semantic embedded feature pyramid | |
| CN103076547B (en) | Method for identifying GIS (Gas Insulated Switchgear) local discharge fault type mode based on support vector machines | |
| CN108919067A (en) | A kind of recognition methods for GIS partial discharge mode | |
| CN114417926A (en) | Power equipment partial discharge pattern recognition method and system based on deep convolution generation countermeasure network | |
| CN116578940A (en) | Bearing fault new type identification and diagnosis method based on mixed depth self-coding | |
| CN110020637A (en) | A kind of analog circuit intermittent fault diagnostic method based on more granularities cascade forest | |
| CN115292538A (en) | Map line element extraction method based on deep learning | |
| CN116934729A (en) | Bridge crack repair trace and secondary crack classification detection method based on image preprocessing and YOLOv7 | |
| CN115112669B (en) | Pavement nondestructive testing identification method based on small sample | |
| CN110610203A (en) | Electric energy quality disturbance classification method based on DWT and extreme learning machine | |
| CN116630989A (en) | Visual fault detection method and system for intelligent ammeter, electronic equipment and storage medium | |
| CN114970601A (en) | Power equipment partial discharge type identification method, equipment and storage medium | |
| CN113919396B (en) | Vibration signal and image characteristic machine tool wear state monitoring method based on semi-supervised learning | |
| Li et al. | Image classification via LZ78 based string kernel: a comparative study | |
| CN115982621A (en) | Rotary machine residual service life prediction method based on time convolution network | |
| CN112560674B (en) | Method and system for detecting sound signal quality | |
| CN111881026A (en) | Variation testing method based on reliability of intelligent traffic model | |
| CN118173101B (en) | Voiceprint-based external damage risk identification method |
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: 20200327 |
|
| RJ01 | Rejection of invention patent application after publication |