CN113449803A - Method, system, equipment and storage medium for distinguishing partial discharge types of different insulation defects - Google Patents
Method, system, equipment and storage medium for distinguishing partial discharge types of different insulation defects Download PDFInfo
- Publication number
- CN113449803A CN113449803A CN202110780818.XA CN202110780818A CN113449803A CN 113449803 A CN113449803 A CN 113449803A CN 202110780818 A CN202110780818 A CN 202110780818A CN 113449803 A CN113449803 A CN 113449803A
- Authority
- CN
- China
- Prior art keywords
- partial discharge
- feature information
- layer
- local
- neural network
- 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/254—Fusion techniques of classification results, e.g. of results related to same input data
-
- 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
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computational Linguistics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Testing Relating To Insulation (AREA)
Abstract
The invention provides a method, a system, equipment and a storage medium for distinguishing partial discharge types of different insulation defects, wherein the method comprises the following steps: extracting local spatial feature information of the partial discharge signal based on multilayer convolution and pooling operations of a convolutional neural network, and reserving time sequence feature information of the local spatial feature information; based on the local spatial feature information, the time series feature information of the local discharge signal is extracted through long-term and short-term memory network storage and analysis; based on the local spatial feature information and the time series feature information, performing feature fusion through a full connection layer, and extracting identification features of all partial discharge signals containing the local spatial feature information and the time series feature information; and outputting a probability value through a softmax layer of the convolutional neural network based on the identification characteristics, and identifying and classifying the type of the partial discharge signal. The invention improves the identification precision of the whole partial discharge image type.
Description
Technical Field
The invention relates to the technical field of image recognition and classification, in particular to a method for distinguishing partial discharge types of different insulation defects.
Background
The convolutional neural network has excellent effect on the partial discharge pattern recognition and classification method of the gas insulated switchgear by using strong automatic feature extraction and classification capability, and the long-time memory network has outstanding advantages in the aspects of data dependency relationship analysis and time dynamics analysis, and provides a new idea for solving the partial discharge pattern recognition problem caused by different insulation defects. Partial discharge is widely recognized as an important factor causing deterioration of insulation of electrical equipment, and is closely related to operational safety and reliability of high-voltage electrical equipment. The method can accurately and automatically identify the discharge type of partial discharge, can timely detect the internal defect and the discharge degree of the insulation, and has great significance for preventing insulation accidents.
At present, a convolutional neural network and a long-term memory network are widely applied to the field of partial discharge image recognition, but the requirements on comprehensive extraction of different characteristics and recognition accuracy of different types of partial discharge images cannot be met.
Disclosure of Invention
The invention provides a method for distinguishing partial discharge types of different insulation defects. And identifying and classifying the partial discharge modes, and combining a convolutional neural network and a long-time and short-time memory network to realize intelligent diagnosis of the partial discharge types so as to improve the identification precision of different types of partial discharge images.
In order to achieve the purpose, the invention adopts the following technical scheme:
a method for distinguishing partial discharge types of different insulation defects comprises the following steps:
extracting local spatial feature information of the partial discharge signal based on multilayer convolution and pooling operations of a convolutional neural network, and reserving time sequence feature information of the local spatial feature information;
based on the local spatial feature information, the time series feature information of the local discharge signal is extracted through long-term and short-term memory network storage and analysis;
based on the local spatial feature information and the time series feature information, performing feature fusion through a full connection layer, and extracting identification features of all partial discharge signals containing the local spatial feature information and the time series feature information;
and outputting a probability value through a softmax layer of the convolutional neural network based on the identification characteristics of all partial discharge signals containing the local space characteristic information and the time series characteristic information, and identifying and classifying the types of the partial discharge signals.
As a further improvement of the invention, the local spatial feature information of the extracted partial discharge signal is automatically extracted by a convolutional neural network to be used as input, and the spatial feature is extracted by the convolutional neural network.
As a further improvement of the present invention, the convolutional neural network is a two-dimensional convolutional neural network, which includes an input layer, a convolutional layer, a pooling layer and a full-link layer; the convolution layer has ten layers, the number of single-layer convolution kernels is respectively 16, 32, 64, 128 and 256, the number of double-layer convolution kernels is the same as that of the previous layer, the pooling layer has five layers, the window size of each layer is 2 x 2, and the step length is 2.
As a further improvement of the invention, the pooling operation of the convolutional neural network adopts maximum pooling in the first four layers and global average pooling in the last layer.
As a further improvement of the present invention, the long-short term memory network comprises four basic components: the device comprises a unit, an input gate, an output gate and a forgetting gate; the number of layers is set to 2 and the number of cells is 128, 64 respectively.
As a further improvement of the present invention, the partial discharge signal types include a metal tip defect, an air gap defect in an insulator, a floating electrode defect, and a free metal particle defect.
A system for differentiating between different insulation defect partial discharge types, comprising:
the first feature extraction module is used for extracting local spatial feature information of the partial discharge signal based on multilayer convolution and pooling operations of a convolutional neural network and reserving time sequence feature information of the local spatial feature information;
the second characteristic extraction module is used for storing and analyzing the local spatial characteristic information through a long-term and short-term memory network and extracting the time series characteristic information of the local discharge signal;
the characteristic fusion module is used for carrying out characteristic fusion through the full connection layer based on the local space characteristic information and the time sequence characteristic information and extracting the identification characteristics of all partial discharge signals containing the local space characteristic information and the time sequence characteristic information;
and the identification feature module is used for outputting a probability value through a softmax layer of the convolutional neural network based on the identification features of all partial discharge signals containing the local space feature information and the time series feature information, and identifying and classifying the types of the partial discharge signals.
An electronic device comprising a memory, a processor and a computer program stored in said memory and executable on said processor, said processor implementing the steps of said different insulation defect partial discharge type discrimination method when executing said computer program.
A computer-readable storage medium, which stores a computer program that, when being executed by a processor, carries out the steps of the method for distinguishing between different insulation defect partial discharge types.
The invention has the beneficial effects that:
the invention provides a method for distinguishing partial discharge types of different insulation defects, which comprises the steps of extracting partial spatial feature information of partial discharge voltage signals through convolution and pooling operation, obtaining the spatial feature of partial discharge signals with higher levels, retaining the sequence feature of data, storing and analyzing the previous voltage signal information through a long-term and short-term memory network, realizing modeling of data time dependence, extracting identification features of all partial discharge signals containing time sequence features, and classifying the partial discharge types through a softmax layer of a convolutional neural network. The method for distinguishing the partial discharge types of different insulation defects can integrate the advantages that a convolutional neural network is good at mining and extracting the spatial characteristics of a partial discharge map and a long-short term memory network is good at mining time sequence characteristic information, can effectively extract and utilize the space-time characteristics of partial discharge signals, and enhances the generalization capability of a model by using a long-short term memory network gate structure, thereby improving the identification precision of the whole partial discharge image type.
Drawings
In order to more clearly illustrate the technical solution of the present invention, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious to those skilled in the art that other drawings can be obtained based on these drawings without creative efforts.
FIG. 1 is a flow chart of a method of distinguishing between partial discharge types caused by different insulation defects.
FIG. 2 is a block diagram of an algorithm for distinguishing between partial discharge types caused by different insulation defects.
FIG. 3 is a flow chart of a method for distinguishing partial discharge types of different insulation defects according to a preferred embodiment of the present invention;
FIG. 4 is a schematic diagram of a system for differentiating partial discharge types of different insulation defects according to a preferred embodiment of the present invention;
fig. 5 is a schematic structural diagram of an electronic device according to a preferred embodiment of the invention.
Detailed Description
Reference will now be made in detail to embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following examples do not represent all embodiments consistent with the present invention. But merely as exemplifications of systems and methods consistent with certain aspects of the invention, as detailed in the claims.
In long-term operation of various large electrical devices, partial discharge is one of the main causes of insulation deterioration. The type of the partial discharge can be judged by detecting the state of the partial discharge of the electrical equipment on line, so that the type of the insulation defect can be diagnosed, and the maintenance can be carried out in time to ensure the normal operation of the electrical equipment. Since partial discharge is closely related to insulation defects of the device, the types of partial discharge caused by different insulation defects are different. By identifying the type of partial discharge in a particular way (e.g. in combination with a convolutional neural network and a long-short term memory network), the nature and severity of the occurring insulation defect can be diagnosed. Then, the insulation defect is located and evaluated for state. At present, a convolutional neural network and a long-term memory network are widely applied to the field of partial discharge image recognition, the method is simple to operate and rapid in recognition, but the requirements on comprehensive extraction of different features and recognition accuracy of different types of partial discharge images cannot be met.
As shown in fig. 3, the present invention provides a method for distinguishing partial discharge types caused by different insulation defects, comprising the following steps:
extracting local spatial feature information of the partial discharge signal based on multilayer convolution and pooling operations of a convolutional neural network, and reserving time sequence feature information of the local spatial feature information;
based on the local spatial feature information, the time series feature information of the local discharge signal is extracted through long-term and short-term memory network storage and analysis;
based on the local spatial feature information and the time series feature information, performing feature fusion through a full connection layer, and extracting identification features of all partial discharge signals containing the local spatial feature information and the time series feature information;
and outputting a probability value through a softmax layer of the convolutional neural network based on the identification characteristics of all partial discharge signals containing the local space characteristic information and the time series characteristic information, and identifying and classifying the types of the partial discharge signals.
The method comprises the steps of automatically extracting features through a convolutional neural network to serve as input, extracting spatial features through the convolutional neural network, extracting output of the convolutional neural network through a long-term memory network to serve as input, and extracting time sequence features through the long-term memory network.
The two-dimensional convolutional neural network has ten layers of convolutional layers, the number of single-layer convolutional kernels is 16, 32, 64, 128 and 256 respectively, the number of double-layer convolutional kernels is the same as that of the former layer, the number of pooling layers is five, the window size of each layer is 2 x 2, the step length is 2, and the local space characteristics of local discharge signals are extracted.
And in the convolutional neural network pooling operation, the first four layers adopt maximum pooling, the last layer adopts global average pooling, and the local spatial features of the partial discharge signals are extracted.
The number of layers of the long-term and short-term memory network is set to be 2, the number of units is 128 and 64 respectively, and the time sequence characteristics of the partial discharge signals are extracted.
According to the technical scheme, the method for distinguishing the partial discharge types of the different insulation defects comprises the steps of extracting the local spatial feature information of the partial discharge voltage signals through convolution and pooling, obtaining the spatial feature of the partial discharge signals of a higher level, retaining the sequence feature of data, storing and analyzing the previous voltage signal information through a long-term and short-term memory network, achieving modeling of data time dependency, extracting the identification features of all the partial discharge signals containing time sequence features, and classifying the partial discharge types through a softmax layer of a convolutional neural network. The invention provides a method for distinguishing partial discharge types of different insulation defects, which can integrate the advantages that a convolutional neural network is good at mining and extracting the spatial characteristics of a partial discharge map and a long-short term memory network is good at mining time sequence characteristic information, can effectively extract and utilize the space-time characteristics of partial discharge signals, and enhances the generalization capability of a model by using a long-short term memory network gate structure, thereby improving the identification precision of the whole partial discharge image type.
Referring to fig. 1, the method specifically includes:
s1: extracting local spatial feature information of the partial discharge signal based on multilayer convolution and pooling operations of a convolutional neural network, and reserving time sequence feature information of the local spatial feature information;
it should be noted that a Convolutional Neural Network (CNN) is an efficient pattern recognition method developed in recent years, and is widely used in the partial discharge pattern recognition classification of gas insulated switchgear due to its strong automatic feature capture capability. Generally, a CNN is mainly composed of an input layer, a convolutional layer, a pooling layer, and a full connection layer. Local connections and parameter sharing in the convolutional neural network reduce the number of parameters, greatly reduce training complexity, and mitigate overfitting. Meanwhile, the tolerance of the convolution network to translation is also given by weight sharing, and the down sampling in the pooling layer further reduces the output parameter number, gives the tolerance of the model to slight deformation, and improves the generalization capability of the model.
In addition, the CNN can adaptively perform feature extraction and data processing, can reduce dimensionality by convolution and merging, and has better generalization capability than conventional features. In partial discharge fault diagnosis, the signal is typically converted to a time domain or time frequency graph and then processed by a 2-dimensional CNN. However, the two-dimensional convolution operation can only extract spatial features, and neglects temporal features of signals, thereby resulting in poor performance of the above model under complex interference. The features are convolutely extracted along the time axis of the signal using a two-dimensional convolution operation taking into account the characteristics of the partial discharge signal, thereby preserving the temporal features while ensuring feature extraction.
S2: and based on the local spatial characteristic information, the time series characteristic information of the local discharge signal is extracted through long-term and short-term memory network storage and analysis.
It should be noted that the long and short-term memory network (LSTM) can store past data in its storage unit, which is very effective for time series data prediction. The use of an LSTM network may alleviate gradient extinction and explosion problems in the recurrent neural network. The LSTM network contains four basic components: the unit, input gate, output gate and forget the gate. Information is transmitted by the units at random time intervals. The input gate determines how much of the network input at the current time is saved to the cell state. How much of the output gate control unit state is output to the current output value of the LSTM. The forgetting gate determines how much the state of the cell at the previous time remains to the current time. LSTM can learn features from sequence data, with unique advantages for processing data in time series.
S3: based on the local spatial feature information and the time series feature information, performing feature fusion through a full connection layer, and extracting identification features of all partial discharge signals containing the local spatial feature information and the time series feature information;
in S3, CNN has a unique advantage in extracting spatial features, and LSTM has a unique advantage in extracting temporal features. However, CNN or LSTM feature extraction alone is poor for a large number of sample signals of partial discharges and requires a large amount of computation time to obtain satisfactory results. Accordingly, the present invention provides a CNN-LSTM hybrid network. First, the convolutional and pooling layers are used to extract spatial features and reduce the dimensionality of the data. The temporal characteristics of the data are then further extracted using the LSTM network.
S4: based on the CNN-LSTM hybrid network, based on the identification characteristics of all partial discharge signals containing the local space characteristic information and the time series characteristic information, a probability value is output through a softmax layer of a convolutional neural network, and the types of the partial discharge signals are identified and classified.
In S4, the CNN-LSTM hybrid network includes a convolutional neural network module, a long-term memory module, and a feature fusion module. The partial discharge signal types comprise metal tip defects, air gap defects in insulators, suspended electrode defects, free metal particle defects and the like.
As shown in fig. 2, first, a single input model is built using four partial discharge time-voltage data sets. CNN employs stacked convolution and pooling operations with very small convolution kernels, which increases the depth of the network model and enhances feature extraction. And then, converting the characteristic value into a characteristic vector, and effectively extracting time sequence characteristics by using an LSTM gate structure to improve the generalization capability of the model. And finally, performing feature fusion on the features of the front layer through the full connection layer, classifying data through the softmax layer, and outputting a probability value to realize feature identification of the partial discharge signal.
The partial discharge pattern recognition device is configured to: an intel 5 processor (2.5GHz), a memory of 16GB, a system of Win10, a running environment of tenserflow ═ 2.0.0, keras ═ 2.3.0, python ═ 3.7, and a typical partial discharge defect voltage signal diagram of class 4 is identified.
A metal tip defect map 62, an air gap defect map 63 in an insulator, a suspended electrode defect map 64 and free metal particles 63 are used, and 252 maps are used as data sets. Wherein 80% of the data samples were used for training and 20% were used for testing, i.e. 201 samples were trained and 51 samples were validated. During model compilation, the learning rate is set to 0.0001, epochs to 65, and batch _ size to 32.
Finally, the provided CNN-LSTM model has the highest identification accuracy rate of air gap, suspension electrode and free metal particle defects in the insulator, reaching 100 percent, and the lowest identification rate of metal tip defects, but also reaching 92 percent. The recognition rate of the proposed CNN-LSTM model on four partial discharge defects averagely reaches 97.6%, and the CNN-LSTM model has good performance in partial discharge mode recognition.
As shown in fig. 4, another objective of the present invention is to provide a system for distinguishing partial discharge types of different insulation defects, comprising:
the first feature extraction module is used for extracting local spatial feature information of the partial discharge signal based on multilayer convolution and pooling operations of a convolutional neural network and reserving time sequence feature information of the local spatial feature information;
the second characteristic extraction module is used for storing and analyzing the local spatial characteristic information through a long-term and short-term memory network and extracting the time series characteristic information of the local discharge signal;
the characteristic fusion module is used for carrying out characteristic fusion through the full connection layer based on the local space characteristic information and the time sequence characteristic information and extracting the identification characteristics of all partial discharge signals containing the local space characteristic information and the time sequence characteristic information;
and the identification feature module is used for outputting a probability value through a softmax layer of the convolutional neural network based on the identification features of all partial discharge signals containing the local space feature information and the time series feature information, and identifying and classifying the types of the partial discharge signals.
A third object of the present invention is to provide an electronic device, as shown in fig. 5, including a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method for distinguishing different insulation defect partial discharge types when executing the computer program.
The method for distinguishing the partial discharge types of the different insulation defects comprises the following steps:
extracting local spatial feature information of the partial discharge signal based on multilayer convolution and pooling operations of a convolutional neural network, and reserving time sequence feature information of the local spatial feature information;
based on the local spatial feature information, the time series feature information of the local discharge signal is extracted through long-term and short-term memory network storage and analysis;
based on the local spatial feature information and the time series feature information, performing feature fusion through a full connection layer, and extracting identification features of all partial discharge signals containing the local spatial feature information and the time series feature information;
and outputting a probability value through a softmax layer of the convolutional neural network based on the identification characteristics of all partial discharge signals containing the local space characteristic information and the time series characteristic information, and identifying and classifying the types of the partial discharge signals.
A fourth object of the present invention is to provide a computer-readable storage medium, which stores a computer program that, when being executed by a processor, implements the steps of the method for distinguishing between different insulation defect partial discharge types.
The method for distinguishing the partial discharge types of the different insulation defects comprises the following steps:
extracting local spatial feature information of the partial discharge signal based on multilayer convolution and pooling operations of a convolutional neural network, and reserving time sequence feature information of the local spatial feature information;
based on the local spatial feature information, the time series feature information of the local discharge signal is extracted through long-term and short-term memory network storage and analysis;
based on the local spatial feature information and the time series feature information, performing feature fusion through a full connection layer, and extracting identification features of all partial discharge signals containing the local spatial feature information and the time series feature information;
and outputting a probability value through a softmax layer of the convolutional neural network based on the identification characteristics of all partial discharge signals containing the local space characteristic information and the time series characteristic information, and identifying and classifying the types of the partial discharge signals.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While there have been shown and described what are at present considered the fundamental principles and essential features of the invention and its advantages, it will be apparent to those skilled in the art that the invention is not limited to the details of the foregoing exemplary embodiments, but is capable of other specific forms without departing from the spirit or essential characteristics thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.
Furthermore, it should be understood that although the present description refers to embodiments, not every embodiment may contain only a single embodiment, and such description is for clarity only, and those skilled in the art should integrate the description, and the embodiments may be combined as appropriate to form other embodiments understood by those skilled in the art.
The embodiments of the present invention are described in detail, and the embodiments are only examples of the general inventive concept, and should not be construed as limiting the scope of the present invention. Any other embodiments extended by the solution according to the invention without inventive step will be within the scope of protection of the invention for a person skilled in the art.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting the same, and although the present invention is described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications and equivalents may be made to the embodiments of the invention without departing from the spirit and scope of the invention, which is to be covered by the claims.
Claims (9)
1. A method for distinguishing partial discharge types of different insulation defects is characterized by comprising the following steps:
extracting local spatial feature information of the partial discharge signal based on multilayer convolution and pooling operations of a convolutional neural network, and reserving time sequence feature information of the local spatial feature information;
based on the local spatial feature information, the time series feature information of the local discharge signal is extracted through long-term and short-term memory network storage and analysis;
based on the local spatial feature information and the time series feature information, performing feature fusion through a full connection layer, and extracting identification features of all partial discharge signals containing the local spatial feature information and the time series feature information;
and outputting a probability value through a softmax layer of the convolutional neural network based on the identification characteristics of all partial discharge signals containing the local space characteristic information and the time series characteristic information, and identifying and classifying the types of the partial discharge signals.
2. The method of claim 1,
and the local spatial feature information of the extracted partial discharge signal is automatically extracted as input through a convolutional neural network, and spatial features are extracted by utilizing the convolutional neural network.
3. The method of claim 1,
the convolutional neural network is a two-dimensional convolutional neural network and comprises an input layer, a convolutional layer, a pooling layer and a full-connection layer; the convolution layer has ten layers, the number of single-layer convolution kernels is respectively 16, 32, 64, 128 and 256, the number of double-layer convolution kernels is the same as that of the previous layer, the pooling layer has five layers, the window size of each layer is 2 x 2, and the step length is 2.
4. The method of claim 1,
the pooling operation of the convolutional neural network is implemented by adopting maximum pooling in the first four layers and global average pooling in the last layer.
5. The method of claim 1,
the long-short term memory network comprises four basic components: the device comprises a unit, an input gate, an output gate and a forgetting gate; the number of layers is set to 2 and the number of cells is 128, 64 respectively.
6. The method of claim 1,
the partial discharge signal types include metal tip defects, air gap defects in insulators, suspended electrode defects, and free metal particle defects.
7. A system for differentiating between partial discharge types of different insulation defects, comprising:
the first feature extraction module is used for extracting local spatial feature information of the partial discharge signal based on multilayer convolution and pooling operations of a convolutional neural network and reserving time sequence feature information of the local spatial feature information;
the second characteristic extraction module is used for storing and analyzing the local spatial characteristic information through a long-term and short-term memory network and extracting the time series characteristic information of the local discharge signal;
the characteristic fusion module is used for carrying out characteristic fusion through the full connection layer based on the local space characteristic information and the time sequence characteristic information and extracting the identification characteristics of all partial discharge signals containing the local space characteristic information and the time sequence characteristic information;
and the identification feature module is used for outputting a probability value through a softmax layer of the convolutional neural network based on the identification features of all partial discharge signals containing the local space feature information and the time series feature information, and identifying and classifying the types of the partial discharge signals.
8. An electronic device comprising a memory, a processor and a computer program stored in said memory and executable on said processor, said processor implementing the steps of the method for distinguishing between different insulation defect partial discharge types according to any one of claims 1 to 6 when executing said computer program.
9. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the steps of the method for distinguishing between different insulation defect partial discharge types according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110780818.XA CN113449803A (en) | 2021-07-09 | 2021-07-09 | Method, system, equipment and storage medium for distinguishing partial discharge types of different insulation defects |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110780818.XA CN113449803A (en) | 2021-07-09 | 2021-07-09 | Method, system, equipment and storage medium for distinguishing partial discharge types of different insulation defects |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113449803A true CN113449803A (en) | 2021-09-28 |
Family
ID=77815919
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110780818.XA Pending CN113449803A (en) | 2021-07-09 | 2021-07-09 | Method, system, equipment and storage medium for distinguishing partial discharge types of different insulation defects |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113449803A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115500807A (en) * | 2022-09-20 | 2022-12-23 | 山东大学 | Arrhythmia classification detection method and system based on small convolutional neural network |
| CN117849560A (en) * | 2024-03-07 | 2024-04-09 | 南京中鑫智电科技有限公司 | Valve side sleeve insulation monitoring method and system combining end screen voltage and partial discharge |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109932627A (en) * | 2019-04-30 | 2019-06-25 | 国网山东省电力公司烟台供电公司 | A kind of localization method and system of GIS pressure test shelf depreciation |
| CN112034310A (en) * | 2020-07-31 | 2020-12-04 | 国网山东省电力公司东营供电公司 | Partial discharge defect diagnosis method and system for combined electrical appliance |
| WO2021057328A1 (en) * | 2019-09-24 | 2021-04-01 | 上海数创医疗科技有限公司 | St segment classification convolutional neural network based on feature selection and method for using same |
-
2021
- 2021-07-09 CN CN202110780818.XA patent/CN113449803A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109932627A (en) * | 2019-04-30 | 2019-06-25 | 国网山东省电力公司烟台供电公司 | A kind of localization method and system of GIS pressure test shelf depreciation |
| WO2021057328A1 (en) * | 2019-09-24 | 2021-04-01 | 上海数创医疗科技有限公司 | St segment classification convolutional neural network based on feature selection and method for using same |
| CN112034310A (en) * | 2020-07-31 | 2020-12-04 | 国网山东省电力公司东营供电公司 | Partial discharge defect diagnosis method and system for combined electrical appliance |
Non-Patent Citations (2)
| Title |
|---|
| TINGLIANG LIU: "《GIS Partial Discharge Pattern Recognition Based on a Novel Convolutional Neural Networks and Long Short-Term Memory》", 《ENTROPY》, pages 1 - 4 * |
| 周秀;朱洪波;马云龙;刘威峰;高博;田天;罗艳;李秀广;何宁辉;: "基于深度学习的变压器局部放电模式识别研究", 高压电器, no. 12 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115500807A (en) * | 2022-09-20 | 2022-12-23 | 山东大学 | Arrhythmia classification detection method and system based on small convolutional neural network |
| CN117849560A (en) * | 2024-03-07 | 2024-04-09 | 南京中鑫智电科技有限公司 | Valve side sleeve insulation monitoring method and system combining end screen voltage and partial discharge |
| CN117849560B (en) * | 2024-03-07 | 2024-05-28 | 南京中鑫智电科技有限公司 | Valve side sleeve insulation monitoring method and system combining end screen voltage and partial discharge |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Do et al. | Convolutional-neural-network-based partial discharge diagnosis for power transformer using UHF sensor | |
| CN110991311B (en) | Target detection method based on dense connection deep network | |
| CN107292886B (en) | Target object intrusion detection method and device based on grid division and neural network | |
| CN113449803A (en) | Method, system, equipment and storage medium for distinguishing partial discharge types of different insulation defects | |
| CN111044570A (en) | Defect identification and early warning method and device for power equipment and computer equipment | |
| CN111695620A (en) | Method and system for detecting and correcting abnormal data of time sequence of power system | |
| CN110362907A (en) | Based on the identification of SSD neural network transmission line of electricity target defect and diagnostic method | |
| Ma et al. | Detection method of insulator based on faster R-CNN | |
| CN116572747B (en) | Battery fault detection method, device, computer equipment and storage medium | |
| CN116167010A (en) | Rapid identification method for abnormal events of power system with intelligent transfer learning capability | |
| CN112633315A (en) | Electric power system disturbance classification method | |
| CN113203914A (en) | Underground cable early fault detection and identification method based on DAE-CNN | |
| CN112950576A (en) | Power transmission line defect intelligent identification method and system based on deep learning | |
| CN115563477A (en) | Harmonic data identification method and device, computer equipment and storage medium | |
| CN117523218A (en) | Label generation, training of image classification model and image classification method and device | |
| CN112686333A (en) | Switch cabinet partial discharge mode identification method based on depth subdomain adaptive migration network | |
| CN115563500A (en) | Power distribution equipment partial discharge mode identification method, device and system based on data enhancement technology | |
| CN115239971A (en) | GIS partial discharge type recognition model training method, recognition method and system | |
| CN115687897A (en) | High-voltage cable aluminum sheath corrosion damage identification method, system, equipment and medium based on semi-supervised learning | |
| CN113762007B (en) | Abnormal behavior detection method based on appearance and action feature double prediction | |
| CN111666819B (en) | High-precision video abnormal event detection method integrating multivariate information | |
| CN114417977A (en) | Fault diagnosis method and system based on transformer data sample expansion | |
| Wu et al. | Network Construction for Bearing Fault Diagnosis Based on Double Attention Mechanism | |
| CN117292331B (en) | Complex foreign matter detection system and method based on deep learning | |
| CN116224000A (en) | Transformer partial discharge type identification method and equipment based on convolutional neural network |
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 |