Partial discharge type identification method and device

By combining the time domain feature extraction network and the frequency domain feature extraction network, the problem of local discharge signal recognition accuracy and low recognition rate is solved, and more efficient signal feature extraction and fault category determination is achieved, which improves the comprehensiveness of signal processing and Validity.

CN119001359APending Publication Date: 2024-11-22ZHONGTIAN ELECTRIC POWER OPTICAL CABLES CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411139183.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-19
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

The prior art has low accuracy of the effective signal when processing local discharge signals, and different types of local discharges may exhibit similar image characteristics in the spectrum diagram, resulting in a lower recognition rate.

Method used

The time domain feature extraction network and frequency domain feature extraction network are used to extract the time domain features and frequency domain features of local discharge signals through the time domain convolution network, spatial attention network, class marker vector splicing layer and position encoding embedding layer, and the time domain features and frequency domain features of local discharge signals are extracted, and Combining the full connection layer and the softmax layer determine the fault category.

Benefits of technology

It improves the extraction efficiency and accuracy of local discharge signals, reduces the impact of interference between signals, and can fully understand signal characteristics from multiple angles, which enhances the comprehensiveness and effectiveness of signal processing and analysis.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119001359A_ABST
    Figure CN119001359A_ABST
Patent Text Reader

Abstract

The invention provides a partial discharge type identification method and device, and is applied to the technical field of signal processing. The method comprises the following steps: collecting current pulse data in a cable; determining a time domain signal and a frequency spectrum image of the current pulse data; inputting the time domain signal into a time domain feature extraction network to obtain time domain feature information of the partial discharge signal; inputting the frequency spectrum image into a frequency domain feature extraction network to obtain frequency domain feature information of the partial discharge signal; determining a fault type corresponding to the partial discharge signal according to the time domain feature information and the frequency domain feature information; wherein the time domain feature extraction network comprises a time domain convolution network and a space attention network, and the frequency domain feature extraction network comprises a class mark vector splicing layer and a position coding embedding layer.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of signal processing, and in particular to a method and device for identifying a partial discharge category. Background Art

[0002] Partial discharge is a common fault phenomenon in power equipment, especially in high-voltage equipment. Partial discharge not only leads to a decline in equipment performance, but may also cause serious equipment damage or even safety accidents. In order to ensure the stable operation of the power system, it is crucial to accurately identify and locate partial discharge.

[0003] In the prior art, in the process of processing partial discharge signals, it is not only affected by environmental noise and human operation errors, resulting in low accuracy of extracted effective signals; but also different types of partial discharges may show similar image features in the spectrum diagram, so the recognition rate is low. Summary of the invention

[0004] The present invention provides a partial discharge category recognition method and device, which are used to solve the problem that in the process of processing partial discharge signals in the prior art, not only the accuracy of extracted effective signals is low, but also the recognition rate is low.

[0005] The present invention provides a method for identifying a partial discharge category, comprising: collecting current pulse data in a cable; determining a time domain signal and a spectrum image of the current pulse data; inputting the time domain signal into a time domain feature extraction network to obtain time domain feature information of the partial discharge signal; inputting the spectrum image into a frequency domain feature extraction network to obtain frequency domain feature information of the partial discharge signal; determining a fault category corresponding to the partial discharge signal according to the time domain feature information and the frequency domain feature information; wherein the time domain feature extraction network comprises a time domain convolutional network and a spatial attention network, and the frequency domain feature extraction network comprises a class label vector concatenation layer and a position encoding embedding layer.

[0006] According to a partial discharge category recognition method provided by the present invention, the time domain signal is input into a time domain feature extraction network to obtain time domain feature information of the partial discharge signal, including: inputting the time domain signal into a two-layer time domain convolution network to obtain convolution features; inputting the time domain signal into a spatial attention network to obtain weights of feature points; and weighted processing the convolution features based on the weights of the feature points to obtain the time domain feature information.

[0007] According to a partial discharge category recognition method provided by the present invention, the frequency spectrum image is input into a frequency domain feature extraction network to obtain frequency domain feature information of a partial discharge signal, including: extracting features from the frequency spectrum image, and reshaping the generated feature map and flattening it into a feature vector sequence; splicing the feature vector sequence, the self-learned class label vector of the class label vector splicing layer, and the self-learned position vector of the position coding embedding layer to obtain a splicing vector; and inputting the splicing vector into an encoder network to extract effective features to obtain the frequency domain feature information.

[0008] According to a partial discharge category identification method provided by the present invention, the method of determining the fault category corresponding to the partial discharge signal according to the time domain feature information and the frequency domain feature information comprises: inputting the time domain feature information and the frequency domain feature information into a fully connected layer and a softmax layer to obtain the probability of the partial discharge category; performing weighted summation according to the probability of the partial discharge category, and outputting the fault category corresponding to the partial discharge signal.

[0009] According to a partial discharge category identification method provided by the present invention, the fault category corresponding to the partial discharge signal includes one or more of surface discharge, gap discharge and corona discharge.

[0010] According to a partial discharge category identification method provided by the present invention, after collecting the current pulse data in the cable, the method further includes: performing a preprocessing operation on the current pulse data, and the preprocessing operation includes filtering abnormal data, global filtering processing, and eliminating interference frequencies.

[0011] The present invention also provides a partial discharge category identification device, comprising the following modules: an acquisition module and a processing module; the acquisition module is used to acquire current pulse data in the cable; the processing module is used to determine the time domain signal and spectrum image of the current pulse data; the time domain signal is input into a time domain feature extraction network to obtain time domain feature information of the partial discharge signal; the spectrum image is input into a frequency domain feature extraction network to obtain frequency domain feature information of the partial discharge signal; the fault category corresponding to the partial discharge signal is determined according to the time domain feature information and the frequency domain feature information; wherein the time domain feature extraction network includes a time domain convolutional network and a spatial attention network, and the frequency domain feature extraction network includes a class label vector splicing layer and a position encoding embedding layer.

[0012] According to a partial discharge category recognition device provided by the present invention, the processing module is used to input the time domain signal into a two-layer time domain convolutional network to obtain convolution features; input the time domain signal into a spatial attention network to obtain weights of feature points; and perform weighted processing on the convolution features based on the weights of the feature points to obtain the time domain feature information.

[0013] According to a partial discharge category recognition device provided by the present invention, the processing module is used to extract features from the spectrum image, and reshape the dimension of the generated feature map and flatten it into a feature vector sequence; the feature vector sequence, the self-learned class label vector of the class label vector splicing layer, and the self-learned position vector of the position coding embedding layer are spliced ​​to obtain a spliced ​​vector; the spliced ​​vector is input into an encoder network to extract effective features to obtain the frequency domain feature information.

[0014] According to a partial discharge category identification device provided by the present invention, the processing module is used to input the time domain feature information and the frequency domain feature information into a fully connected layer and a softmax layer to obtain the probability of the partial discharge category; perform weighted summation according to the probability of the partial discharge category, and output the fault category corresponding to the partial discharge signal.

[0015] According to a partial discharge type identification device provided by the present invention, the fault type corresponding to the partial discharge signal includes one or more of surface discharge, gap discharge and corona discharge.

[0016] According to a partial discharge classification identification device provided by the present invention, the processing module is further used to perform a preprocessing operation on the current pulse data, and the preprocessing operation includes filtering abnormal data, global filtering processing, and eliminating interference frequencies.

[0017] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any of the above-mentioned partial discharge category identification methods is implemented.

[0018] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the partial discharge category identification method as described in any one of the above is implemented.

[0019] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method for identifying a partial discharge category as described above is implemented.

[0020] The method and device for identifying the type of partial discharge provided by the present invention can collect current pulse data in a cable; determine the time domain signal and spectrum image of the current pulse data; input the time domain signal into a time domain feature extraction network to obtain time domain feature information of the partial discharge signal; input the spectrum image into a frequency domain feature extraction network to obtain frequency domain feature information of the partial discharge signal; determine the fault type corresponding to the partial discharge signal according to the time domain feature information and the frequency domain feature information; wherein the time domain feature extraction network includes a time domain convolutional network and a spatial attention network, and the frequency domain feature extraction network includes a class label vector concatenation layer and a position encoding embedding layer. Through this scheme, since the time domain characteristic information of the partial discharge signal can be obtained, the transient characteristics and timing information of the partial discharge signal can be effectively captured, and the response capability to the dynamic changes of the signal can be improved; since the frequency domain characteristic information of the partial discharge signal can be obtained, the complex frequency domain signal can be better processed, the influence of interference between signals can be reduced, and the efficiency and accuracy of partial discharge feature extraction can be improved; since the time domain characteristic information and the frequency domain characteristic information can be combined to determine the fault category corresponding to the partial discharge signal, the characteristics of the signal can be fully understood and described from multiple angles, thereby improving the comprehensiveness and effectiveness of signal processing and analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0022] Figure 1 This is one of the flow charts of the partial discharge category identification method provided by the present invention; Figure 2 This is the second flow chart of the partial discharge category identification method provided by the present invention; Figure 3 It is a structural schematic diagram of a partial discharge category identification device provided by the present invention; Figure 4 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this application. Obviously, the described embodiments are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0024] It should be noted that, in the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.

[0025] It should be noted that, in this article, the terms "comprise", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise one..." do not exclude the presence of other identical elements in the process, method, article or device including the element. In addition, it should be pointed out that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved, for example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.

[0026] In order to facilitate a clear description of the technical solutions of the embodiments of the present application, in the embodiments of the present application, words such as "first" and "second" are used to distinguish between identical or similar items with substantially the same functions and effects. Those skilled in the art can understand that words such as "first" and "second" do not limit the quantity and order of execution.

[0027] The embodiments of the present application describe some exemplary embodiments for the purpose of explanation. It should be understood that the present application can be implemented in other ways that are not specifically shown in the drawings.

[0028] like Figure 1 As shown, the embodiment of the present application provides a partial discharge classification identification method, which can be applied to a partial discharge classification identification device. The partial discharge classification identification method may include S101-S105: S101. A partial discharge category identification device collects current pulse data in a cable.

[0029] The partial discharge category identification device can collect current pulses in the cable through sensors to obtain current pulse data.

[0030] S102: A partial discharge type identification device determines a time domain signal and a spectrum image of the current pulse data.

[0031] Optionally, the partial discharge category identification device may convert the collected current pulse data into a time domain signal, and generate a spectrum image by Fourier transforming the current pulse data.

[0032] Optionally, the partial discharge type identification device may first perform a preprocessing operation on the current pulse data, wherein the preprocessing operation includes filtering abnormal data, global filtering, and eliminating interference frequencies.

[0033] Specifically, the partial discharge category identification device can pre-process the collected current pulse data, delete incomplete or missing abnormal data, perform global filtering on the data, and use a wavelet transform filter to remove interference frequencies to ensure the smoothness of the signal.

[0034] S103: The partial discharge type identification device inputs the time domain signal into a time domain feature extraction network to obtain time domain feature information of the partial discharge signal.

[0035] Among them, the time domain feature extraction network includes a time domain convolutional network and a spatial attention network.

[0036] Optionally, the step of inputting the time domain signal into a time domain feature extraction network to obtain time domain feature information of the local discharge signal includes: inputting the time domain signal into a two-layer time domain convolutional network to obtain convolution features; inputting the time domain signal into a spatial attention network to obtain weights of feature points; and weighting the convolution features based on the weights of the feature points to obtain the time domain feature information.

[0037] Specifically, Figure 2 As shown, the partial discharge category recognition device can first input the time domain signal into the time domain convolutional network, capture the long-term dependency in the time series data through the causal hole convolution layer, and then normalize the weight parameter information generated by the convolution, and pass the Relu activation function to reduce the instability of the training process. At the same time, the partial discharge category recognition device can also input the time domain signal into the spatial attention network, perform maximum pooling and average pooling on all channels of each feature point after different convolutions, stack the pooled results and perform convolution fusion to extract and enhance the information of the feature points, and then generate the weight of each feature point through the Sigmoid activation function. The final weight is used to weight the feature information output by the two-layer time domain convolutional network to generate a feature representation of the time domain information, that is, the time domain feature information.

[0038] By introducing the spatial attention mechanism, not only the ability of the time domain feature extraction network in processing long-distance dependencies and capturing global information is expanded, but also the network's adaptability to complex signals and environmental changes is enhanced.

[0039] S104: The partial discharge type identification device inputs the spectrum image into a frequency domain feature extraction network to obtain frequency domain feature information of the partial discharge signal.

[0040] Wherein, the frequency domain feature extraction network includes a class label vector concatenation layer and a position encoding embedding layer.

[0041] Optionally, the inputting the spectrum image into a frequency domain feature extraction network to obtain frequency domain feature information of the local discharge signal includes: extracting features from the spectrum image, and reshaping the generated feature map and flattening it into a feature vector sequence; splicing the feature vector sequence, the self-learned class label vector of the class label vector splicing layer, and the self-learned position vector of the position coding embedding layer to obtain a splicing vector; and inputting the splicing vector into an encoder network to perform effective feature extraction to obtain the frequency domain feature information.

[0042] Specifically, Figure 2 As shown, the local discharge category recognition device can extract features from the spectrum image through convolution, reshape the dimension of the generated feature map and flatten it into a feature vector sequence. Then, the generated feature vector is spliced ​​with the self-learning class label vector and added to the self-learning position vector to obtain a spliced ​​vector, which will be used as the input of the encoder network. The spliced ​​vector with the position vector added is input into the encoder network and processed by a multi-encoder network to extract and learn effective feature representations. The encoded spliced ​​vector first passes through a normalization layer to normalize all channels of each sample, so that the input distribution of each layer is more stable, and then passes through multiple groups of self-attention layers, so that the model is more focused on the information at different positions of the input sequence. Finally, the frequency domain feature information is obtained through a feedforward fully connected layer and a normalization layer. The feedforward fully connected layer and the normalization layer can enhance the model's expressive power and the ability to capture local features in the sequence, and map the input data into a feature representation with more semantic information; at the same time, the encoder network uses residual connections to introduce direct paths across layers, which can solve the gradient problem in deep neural network training.

[0043] Optionally, the structure of the encoder network consists of a normalization layer, a self-attention layer, a feed-forward fully connected layer, and a residual structure, etc. The feed-forward fully connected layer consists of a fully connected layer, a GELU activation function layer, and a Dropout layer.

[0044] Spectrograms usually have highly complex features and patterns. By adopting an encoder structure to process the frequency domain part, the network can automatically learn and capture these features, thereby improving the robustness of the model.

[0045] S105. The partial discharge type identification device determines a fault type corresponding to the partial discharge signal according to the time domain characteristic information and the frequency domain characteristic information.

[0046] Optionally, determining the fault category corresponding to the local discharge signal according to the time domain feature information and the frequency domain feature information includes: inputting the time domain feature information and the frequency domain feature information into a fully connected layer and a softmax layer to obtain the probability of the local discharge category; performing weighted summation according to the probability of the local discharge category, and outputting the fault category corresponding to the local discharge signal.

[0047] Optionally, the fault category corresponding to the partial discharge signal includes one or more of surface discharge, gap discharge and corona discharge.

[0048] It should be noted that the present application also provides a classification recognition network, including the time domain feature extraction network, frequency domain feature extraction network, fully connected layer and softmax layer in the above-mentioned local discharge category recognition method. The local discharge category recognition device can collect different types of local discharge events through sensor equipment, including fault categories such as surface discharge, gap discharge and corona discharge, and then classify the processed data according to different fault categories, and mark the different fault categories as output labels. Construct a data set according to 7:3, 70% as a training set and 30% as a test set. Finally, the cross entropy loss function F is used to measure the difference between the model prediction result and the true label, and iterative calculation is performed until the model converges.

[0049] The cross entropy loss function F can be expressed as: ; in, is the true label, is the model prediction probability, is the batch size.

[0050] In the embodiment of the present application, since the time domain characteristic information of the local discharge signal can be obtained, the transient characteristics and timing information of the local discharge signal can be effectively captured, and the response capability to the dynamic changes of the signal can be improved; since the frequency domain characteristic information of the local discharge signal can be obtained, the complex frequency domain signal can be better processed, the influence of interference between signals can be reduced, and the efficiency and accuracy of extracting the local discharge characteristics can be improved; since the time domain characteristic information and the frequency domain characteristic information can be combined to determine the fault category corresponding to the local discharge signal, the characteristics of the signal can be fully understood and described from multiple angles, thereby improving the comprehensiveness and effectiveness of signal processing and analysis.

[0051] The above mainly introduces the solution provided by the embodiment of the present application from the perspective of the method. In order to realize the above functions, it includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should easily realize that, in combination with the units and algorithm steps of each example described in the embodiment disclosed herein, the embodiment of the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0052] It should be noted that the device in the embodiment of the present application includes a virtual device and a physical device. The virtual device may be a partial discharge category identification device, and the physical device may include an electronic device, a computer storage medium, and a computer program product.

[0053] The partial discharge classification identification method provided in the embodiment of the present application can be executed by a partial discharge classification identification device or a control module for partial discharge classification identification in the partial discharge classification identification device. In the embodiment of the present application, the partial discharge classification identification method performed by the partial discharge classification identification device is taken as an example to illustrate the partial discharge classification identification device provided in the embodiment of the present application.

[0054] It should be noted that the embodiment of the present application can divide the functional modules of the partial discharge category identification device according to the above method example. For example, each functional module can be divided according to each function, or two or more functions can be integrated into one processing module. The above integrated module can be implemented in the form of hardware or in the form of software functional modules. Optionally, the division of modules in the embodiment of the present application is schematic and is only a logical function division. There may be other division methods in actual implementation.

[0055] like Figure 3As shown, an embodiment of the present application provides a partial discharge category identification device 300. The partial discharge category identification device 300 includes: an acquisition module 301 and a processing module 302; the acquisition module 301 is used to collect current pulse data in the cable; the processing module 302 is used to determine the time domain signal and spectrum image of the current pulse data; the time domain signal is input into the time domain feature extraction network to obtain the time domain feature information of the partial discharge signal; the spectrum image is input into the frequency domain feature extraction network to obtain the frequency domain feature information of the partial discharge signal; the fault category corresponding to the partial discharge signal is determined according to the time domain feature information and the frequency domain feature information; wherein the time domain feature extraction network includes a time domain convolutional network and a spatial attention network, and the frequency domain feature extraction network includes a class label vector concatenation layer and a position encoding embedding layer.

[0056] Optionally, the processing module 302 is used to input the time domain signal into a two-layer time domain convolutional network to obtain convolution features; input the time domain signal into a spatial attention network to obtain weights of feature points; and perform weighted processing on the convolution features based on the weights of the feature points to obtain the time domain feature information.

[0057] Optionally, the processing module 302 is used to extract features from the spectrum image, and reshape the generated feature map and flatten it into a feature vector sequence; splice the feature vector sequence, the self-learned class label vector of the class label vector splicing layer, and the self-learned position vector of the position coding embedding layer to obtain a spliced ​​vector; input the spliced ​​vector into the encoder network for effective feature extraction to obtain the frequency domain feature information.

[0058] Optionally, the processing module 302 is used to input the time domain feature information and the frequency domain feature information into a fully connected layer and a softmax layer to obtain the probability of a partial discharge category; perform weighted summation according to the probability of the partial discharge category, and output a fault category corresponding to the partial discharge signal.

[0059] Optionally, the fault category corresponding to the partial discharge signal includes one or more of surface discharge, gap discharge and corona discharge.

[0060] Optionally, the processing module 302 is further used to perform a preprocessing operation on the current pulse data, wherein the preprocessing operation includes filtering abnormal data, global filtering, and eliminating interference frequencies.

[0061] In the embodiment of the present application, since the time domain characteristic information of the local discharge signal can be obtained, the transient characteristics and timing information of the local discharge signal can be effectively captured, and the response capability to the dynamic changes of the signal can be improved; since the frequency domain characteristic information of the local discharge signal can be obtained, the complex frequency domain signal can be better processed, the influence of interference between signals can be reduced, and the efficiency and accuracy of extracting the local discharge characteristics can be improved; since the time domain characteristic information and the frequency domain characteristic information can be combined to determine the fault category corresponding to the local discharge signal, the characteristics of the signal can be fully understood and described from multiple angles, thereby improving the comprehensiveness and effectiveness of signal processing and analysis.

[0062] Figure 4 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 4 As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430 and a communication bus 440, wherein the processor 410, the communication interface 420 and the memory 430 communicate with each other through the communication bus 440. The processor 410 may call the logic instructions in the memory 430 to execute the partial discharge category recognition method, which includes: collecting current pulse data in the cable; determining the time domain signal and the spectrum image of the current pulse data; inputting the time domain signal into the time domain feature extraction network to obtain the time domain feature information of the partial discharge signal; inputting the spectrum image into the frequency domain feature extraction network to obtain the frequency domain feature information of the partial discharge signal; determining the fault category corresponding to the partial discharge signal according to the time domain feature information and the frequency domain feature information; wherein the time domain feature extraction network includes a time domain convolution network and a spatial attention network, and the frequency domain feature extraction network includes a class label vector concatenation layer and a position encoding embedding layer.

[0063] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0064] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the local discharge category identification method provided by the above methods, which method includes: collecting current pulse data in the cable; determining the time domain signal and spectrum image of the current pulse data; inputting the time domain signal into a time domain feature extraction network to obtain time domain feature information of the local discharge signal; inputting the spectrum image into a frequency domain feature extraction network to obtain frequency domain feature information of the local discharge signal; determining the fault category corresponding to the local discharge signal according to the time domain feature information and the frequency domain feature information; wherein the time domain feature extraction network includes a time domain convolutional network and a spatial attention network, and the frequency domain feature extraction network includes a class label vector concatenation layer and a position encoding embedding layer.

[0065] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the local discharge category identification method provided by the above-mentioned methods, the method comprising: collecting current pulse data in the cable; determining the time domain signal and spectrum image of the current pulse data; inputting the time domain signal into a time domain feature extraction network to obtain time domain feature information of the local discharge signal; inputting the spectrum image into a frequency domain feature extraction network to obtain frequency domain feature information of the local discharge signal; determining the fault category corresponding to the local discharge signal according to the time domain feature information and the frequency domain feature information; wherein the time domain feature extraction network comprises a time domain convolutional network and a spatial attention network, and the frequency domain feature extraction network comprises a class label vector concatenation layer and a position encoding embedding layer.

[0066] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0067] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0068] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying partial discharge types, characterized in that: include: Collect current pulse data in the cable; Determining a time domain signal and a spectrum image of the current pulse data; Inputting the time domain signal into a time domain feature extraction network to obtain time domain feature information of a partial discharge signal; Inputting the spectrum image into a frequency domain feature extraction network to obtain frequency domain feature information of a partial discharge signal; Determine a fault type corresponding to the partial discharge signal according to the time domain characteristic information and the frequency domain characteristic information; Among them, the time domain feature extraction network includes a time domain convolutional network and a spatial attention network, and the frequency domain feature extraction network includes a class label vector concatenation layer and a position encoding embedding layer.

2. The method for identifying partial discharge types according to claim 1, characterized in that: The step of inputting the time domain signal into a time domain feature extraction network to obtain time domain feature information of the partial discharge signal includes: Inputting the time domain signal into a two-layer time domain convolutional network to obtain convolution features; Inputting the time domain signal into the spatial attention network to obtain the weight of the feature point; The convolution feature is weighted based on the weight of the feature point to obtain the time domain feature information.

3. The method for identifying partial discharge types according to claim 1, characterized in that: The step of inputting the spectrum image into a frequency domain feature extraction network to obtain frequency domain feature information of a partial discharge signal includes: Extracting features from the spectrum image, and reshaping and flattening the generated feature map into a feature vector sequence; The feature vector sequence, the self-learning class label vector of the class label vector splicing layer, and the self-learning position vector of the position coding embedding layer are spliced ​​to obtain a splicing vector; The concatenated vector is input into an encoder network to extract effective features, thereby obtaining the frequency domain feature information.

4. The method for identifying partial discharge types according to claim 1, characterized in that: The determining the fault type corresponding to the partial discharge signal according to the time domain characteristic information and the frequency domain characteristic information includes: Inputting the time domain feature information and the frequency domain feature information into a fully connected layer and a softmax layer to obtain the probability of a partial discharge category; A weighted sum is performed according to the probability of the partial discharge category, and the fault category corresponding to the partial discharge signal is output.

5. The method for identifying partial discharge types according to claim 1, characterized in that: The fault category corresponding to the partial discharge signal includes one or more of surface discharge, gap discharge and corona discharge.

6. The method for identifying partial discharge types according to claim 1, characterized in that: After collecting the current pulse data in the cable, the method further includes: The current pulse data is preprocessed, and the preprocessing operation includes filtering abnormal data, global filtering, and eliminating interference frequencies.

7. A partial discharge type identification device, characterized in that: include: Acquisition module and processing module; The acquisition module is used to collect current pulse data in the cable; The processing module is used to determine the time domain signal and spectrum image of the current pulse data; input the time domain signal into the time domain feature extraction network to obtain the time domain feature information of the partial discharge signal; input the spectrum image into the frequency domain feature extraction network to obtain the frequency domain feature information of the partial discharge signal; determine the fault category corresponding to the partial discharge signal according to the time domain feature information and the frequency domain feature information; Among them, the time domain feature extraction network includes a time domain convolutional network and a spatial attention network, and the frequency domain feature extraction network includes a class label vector concatenation layer and a position encoding embedding layer.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the partial discharge category identification method according to any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for identifying a partial discharge category according to any one of claims 1 to 6 is implemented.

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