Transformer fault detection method, device, equipment and medium

Abstract

The embodiment of the invention discloses a transformer fault detection method and device, equipment and a medium. The method comprises the following steps: collecting a parameter set of soluble gas in oil in the transformer; inputting the parameter set of the dissolved gas in oil into a generative adversarial network to obtain target gas parameter characteristics; inputting the target gas parameter features into a feature decision fusion model to obtain target fusion features; and determining the fault type of the transformer according to the target fusion feature. According to the technical scheme, the parameter set of the dissolved gas in the oil is processed through the generative adversarial network model and the feature decision fusion model, so that the fault type of the transformer is determined, and the accuracy and reliability of transformer fault detection are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power equipment, and in particular to a transformer fault detection method, device, equipment and medium. Background Art

[0002] As one of the core equipment of the power system, the operating status of the power transformer directly affects the safety and stability of the entire power grid. The insulation material inside the transformer may age, overheat, and cause various faults during long-term operation.

[0003] In order to detect and deal with these faults in time, dissolved gas analysis (DGA) has become an effective monitoring method. By analyzing the concentration of various gases dissolved in transformer oil, it can be determined whether there is any abnormality inside the transformer.

[0004] Traditional DGA-based fault diagnosis methods mainly rely on rules and standards developed by human experience, such as the IEC60599 three-ratio method. However, this method often has difficulty in accurately distinguishing different types of faults, especially in the early stages of a fault when the gas content does not change much, and the diagnostic results are easily affected by subjective factors. In addition, with the continuous expansion of the scale of power systems and the continuous development of technology, traditional fault diagnosis methods have been unable to meet the needs of modern power systems. Summary of the invention

[0005] The present invention provides a transformer fault detection method, device, equipment and medium, which respectively processes a set of dissolved gas parameters in oil by generating an adversarial network model and a feature decision fusion model, and then determines the fault type of the transformer, thereby improving the accuracy and reliability of transformer fault detection.

[0006] According to one aspect of the present invention, there is provided a transformer fault detection method, comprising:

[0007] Collect the parameter set of dissolved gas in transformer oil;

[0008] Inputting the set of dissolved gas parameters in the oil into a generative adversarial network to obtain target gas parameter characteristics;

[0009] Inputting the target gas parameter characteristics into a feature decision fusion model to obtain a target fusion characteristic;

[0010] The fault type of the transformer is determined according to the target fusion feature.

[0011] Optionally, the generative adversarial network includes multiple convolution kernels;

[0012] The set of dissolved gas parameters in oil is input into the generative adversarial network to obtain the target gas parameter features, including:

[0013] Convolution processing is performed on the set of dissolved gas parameters in the oil by using multiple convolution kernels to obtain a multi-level feature vector;

[0014] The multi-level feature vectors are pooled and normalized to obtain target gas parameter features.

[0015] Optionally, the multiple convolution kernels include a first convolution kernel, a second convolution kernel and a third convolution kernel;

[0016] The set of dissolved gas parameters in oil is convolved by multiple convolution kernels to obtain a multi-level feature vector, including:

[0017] Performing convolution processing on the set of dissolved gas parameters in oil by using the first convolution kernel to obtain a first eigenvector;

[0018] Performing convolution processing on the first eigenvector using the second convolution kernel to obtain a second eigenvector;

[0019] Performing convolution processing on the second eigenvector through the third convolution kernel to obtain a third eigenvector;

[0020] The first feature vector, the second feature vector and the third feature vector are channel-connected to obtain a multi-level feature vector.

[0021] Optionally, the feature decision fusion model includes a feature fusion layer;

[0022] The target gas parameter feature is input into the feature decision fusion model to obtain the target fusion feature, including:

[0023] Performing feature fusion processing on the target gas parameter features through the feature fusion layer to obtain initial fusion feature data;

[0024] The initial fusion feature data and the target gas parameter feature are channel-connected to obtain a target fusion feature.

[0025] Optionally, performing feature fusion processing on the target gas parameter features through the feature fusion layer to obtain initial fusion feature data includes:

[0026] Determining gas variation characteristics of the target gas parameter characteristics;

[0027] Determining first-level gas characteristics and second-level gas characteristics according to the gas change characteristics;

[0028] The first-level gas features and the second-level gas features are feature fused to obtain initial fused feature data.

[0029] Optionally, the first-level gas feature and the second-level gas feature are feature-fused to obtain initial fused feature data, including:

[0030] Performing a scale reconstruction operation on the first-level gas features to obtain reconstructed gas features;

[0031] The reconstructed gas feature is fused with the second-level gas feature to obtain initial fused feature data.

[0032] Optionally, determining the fault type of the transformer according to the target fusion feature includes:

[0033] The target fusion features are processed by a classification function to obtain the fault type of the corresponding transformer.

[0034] According to another aspect of the present invention, there is provided a transformer fault detection device, comprising:

[0035] Data acquisition module, used to collect the set of dissolved gas parameters in transformer oil;

[0036] A gas parameter feature acquisition module, used for inputting the set of dissolved gas parameters in the oil into a generative adversarial network to obtain target gas parameter features;

[0037] A target fusion feature acquisition module is used to input the target gas parameter feature into a feature decision fusion model to obtain a target fusion feature;

[0038] A fault type determination module is used to determine the fault type of the transformer according to the target fusion feature.

[0039] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0040] at least one processor; and

[0041] a memory communicatively connected to the at least one processor; wherein,

[0042] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the transformer fault detection method described in any embodiment of the present invention.

[0043] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the transformer fault detection method described in any embodiment of the present invention when executed.

[0044] The technical solution of the embodiment of the present invention collects a set of parameters of dissolved gas in oil in the transformer; inputs the set of parameters of dissolved gas in oil into a generative adversarial network to obtain target gas parameter features; inputs the target gas parameter features into a feature decision fusion model to obtain target fusion features; and determines the fault type of the transformer according to the target fusion features. This technical solution processes the set of parameters of dissolved gas in oil through a generative adversarial network model and a feature decision fusion model respectively, and then determines the fault type of the transformer, thereby improving the accuracy and reliability of transformer fault detection.

[0045] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0047] Figure 1 is a flow chart of a transformer fault detection method provided according to Embodiment 1 of the present invention;

[0048] Figure 2 This is a structural example diagram of a generative adversarial network provided according to Embodiment 1 of the present invention;

[0049] Figure 3 is a flow chart of a transformer fault detection method provided according to Embodiment 2 of the present invention;

[0050] Figure 4 is a structural schematic diagram of a transformer fault detection device provided according to Embodiment 3 of the present invention;

[0051] Figure 5 It is a structural schematic diagram of an electronic device provided according to Embodiment 4 of the present invention. DETAILED DESCRIPTION

[0052] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0053] It should be noted that the terms "first", "second", "initial" and "target" in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0054] Embodiment 1

[0055] Figure 1 This is a flowchart of a transformer fault detection method provided according to the first embodiment of the present invention. This embodiment is applicable to the case of transformer fault detection. The method can be performed by a transformer fault detection device. The transformer fault detection device can be implemented in the form of hardware and / or software. The transformer fault detection device can be configured in an electronic device with data processing capabilities. Figure 1 As shown, the method includes:

[0056] S110. Collect a set of parameters of dissolved gases in transformer oil.

[0057] Among them, the transformer can be an electric transformer device in the power system. The set of dissolved gas parameters in oil can be understood as a set formed by various gas concentration parameters generated during the operation of the transformer. The set of dissolved gas parameters in oil can include gas parameters such as hydrogen (H2), methane (CH4), ethane (C2H6), ethylene (C2H4), acetylene (C2H2), etc. Specifically, in the operation process of the transformer equipment in this embodiment, due to the effects of electrical, thermal, mechanical and chemical stresses, the insulating materials and oil inside the equipment will gradually age and decompose, generating various gas concentrations such as hydrogen (H2), methane (CH4), ethane (C2H6), ethylene (C2H4), acetylene (C2H2), carbon monoxide (CO) and carbon dioxide (CO2).

[0058] In this embodiment, a multi-parameter coupled oil-immersed transformer intelligent sensor device based on a micro-electromechanical system can be used to collect multiple operating parameters of the transformer and form a set, namely, a set of dissolved gas parameters in oil; wherein different types of gases in the set of dissolved gas parameters in oil and their concentration changes often correspond to different types of faults inside the transformer equipment.

[0059] S120, inputting the parameter set of dissolved gas in oil into a generative adversarial network to obtain target gas parameter characteristics.

[0060] Among them, the generative adversarial network can be a trained neural network model. In this embodiment, the generative adversarial network can be used to perform operations such as feature extraction and feature scale processing on the set of dissolved gas parameters in oil. Specifically, the generative adversarial network can be constructed based on a generative adversarial network structure that integrates feature extraction and feature merging processing. In this embodiment, the generative adversarial network can be composed of a generator G, a discriminator D and a reconstruction network R. The target gas parameter feature can be understood as the fault gas parameter feature obtained by extracting the features of the gas parameters that obviously have faults in the set of dissolved gas parameters in oil through the generative adversarial network. It can be understood that in this embodiment, different concentrations of dissolved gas in oil are set with corresponding normal value ranges. If the concentrations of each gas in the set of dissolved gas parameters in oil are obviously not in the normal value range, it can be considered that it belongs to the gas parameter with faults. The target gas parameter feature can include multiple fault gas parameter features with different features. In this embodiment, the feature extraction layer contained in the generative adversarial network can be used to perform feature extraction processing and feature scale processing on the set of dissolved gas parameters in oil to obtain fault gas parameter features with different features.

[0061] In this embodiment, optionally, the generative adversarial network includes multiple convolution kernels; the set of dissolved gas parameters in oil is input into the generative adversarial network to obtain target gas parameter characteristics, including: convolution processing is performed on the set of dissolved gas parameters in oil through multiple convolution kernels to obtain a multi-level feature vector; pooling and normalization processing is performed on the multi-level feature vector to obtain the target gas parameter characteristics.

[0062] Among them, the multiple convolution kernels may include a set number of convolution kernels, which can be used to perform convolution processing on the input set of dissolved gas parameters in oil. In this embodiment, the generative adversarial network may include multiple convolution kernels of different sizes. The multi-level feature vectors may be feature vectors of different levels obtained after convolution operations are performed on the gas parameters by multiple convolution kernels of different sizes. Among them, the pooling process may be a maximum pooling process performed on the multi-level feature vectors to extract the significant features of the multi-level features. The normalization process may be to scale the multi-level feature vectors according to a certain ratio so that they fall within a specific interval, thereby eliminating the data range differences between the multi-level feature vectors. In this embodiment, the multi-level feature vectors may be subjected to maximum pooling and data normalization and other processing operations to obtain the target gas parameter features.

[0063] Through such a setting in this embodiment, the convolution kernel in the generative adversarial network can be used to perform convolution processing on the set of dissolved gas parameters in the oil, as well as operations such as pooling processing, to obtain multiple levels of fault gas parameter features, effectively merging multiple gas features of different scales to enhance the reliability of the data.

[0064] In this embodiment, optionally, the multiple convolution kernels include a first convolution kernel, a second convolution kernel and a third convolution kernel; the set of dissolved gas parameters in oil is convolved by the multiple convolution kernels to obtain a multi-level feature vector, including: convolving the set of dissolved gas parameters in oil by the first convolution kernel to obtain a first feature vector; convolving the first feature vector by the second convolution kernel to obtain a second feature vector; convolving the second feature vector by the third convolution kernel to obtain a third feature vector; and channel-connecting the first feature vector, the second feature vector and the third feature vector to obtain a multi-level feature vector.

[0065] The first eigenvector may be a eigenvector obtained by convolving each gas parameter in the set of dissolved gas parameters in oil using the first convolution kernel. The second eigenvector may be a eigenvector obtained by convolving the first eigenvector using the second convolution kernel. The third eigenvector may be a eigenvector obtained by convolving the second eigenvector using the third convolution kernel.

[0066] In this embodiment, the first convolution kernel, the second convolution kernel and the third convolution kernel included in the multiple convolution kernels in the generative adversarial network can be convolution kernels with different weights. The gas parameters in the set of dissolved gas parameters in oil are convolved in turn through the convolution kernels of different sizes to obtain feature vectors of different levels, and then the feature vectors of different levels are connected through the channel axis to obtain a multi-level feature vector.

[0067] For example, the structure of the generated adversarial network in this embodiment is shown in the figure below: Figure 2 As shown. The generative adversarial network includes an input layer, a building block 1, a building block 2, a building block 3, a feature merging layer and an output layer. Among them, the building block 1, the building block 2, and the building block 3 respectively contain the first convolution kernel W1, the second convolution kernel W2 and the third convolution kernel W3, and the feature 1, the feature 2 and the feature 3 correspond to the first feature vector F1, the second feature vector F2 and the third feature vector F3 respectively. Specifically, this embodiment uses a plurality of convolution kernels W1, W2, and W3 of different sizes to perform a convolution operation on the oil-soluble gas parameter set I:

[0068] F1=W1*I

[0069] F2=W2*F1

[0070] F3=W3*F2;

[0071] Among them, F1 can be expressed as the first eigenvector, F2 can be expressed as the second eigenvector, F3 can be expressed as the third eigenvector, and W1, W2, and W3 can also be expressed as the weights of the convolution operation.

[0072] Then, the first eigenvector, the second eigenvector, and the third eigenvector are connected through the channel axis to obtain a multi-level eigenvector F:

[0073] F=Concat(F1,F2,F3,axis=Channel);

[0074] Among them, F and Fi represent the feature vectors in different layers respectively, and Concat(·,·,·,axis=Channel) indicates that the connection is performed on the channel axis of the feature.

[0075] Among them, the operations such as maximum pooling and batch normalization are as follows:

[0076] O = MP(ReLU(BN(W4*F)));

[0077] Among them, O is the output of the generative adversarial network, BN represents batch normalization, ReLU represents the activation function, and MP represents maximum pooling.

[0078] In addition, in this embodiment, in order to ensure the size consistency of features of different scales, the step parameter and the padding parameter can be set to adjust the scale. Specifically, a convolution operation with a step parameter of 1 and a padding parameter of 2 can be used to ensure the size consistency of the features.

[0079] Through such a setting in this embodiment, gas features at different levels can be effectively extracted, and gas features at different scales can be fused to obtain multi-level feature vectors, thereby enhancing the quality of generated gas data and ensuring the consistency of data scale.

[0080] S130, inputting the target gas parameter characteristics into the feature decision fusion model to obtain the target fusion characteristics.

[0081] Among them, the feature decision fusion model can be understood as a trained network model, which can be used to perform feature fusion on the extracted gas parameter features of different features to obtain the target fusion feature. The target fusion feature can be understood as the feature obtained by fusing the gas parameter features of multiple different features. Specifically, the different features may include higher-level and lower-level gas parameter features. In this embodiment, the target gas parameter feature can be input into the feature decision fusion model, and the lower-level feature layer can be processed by the convolution operation of the feature decision fusion model to adjust the scale of the feature as a weight, and then the adjusted feature is connected with the higher-level feature, and the feature fusion is further completed through the convolution operation to obtain the target fusion feature.

[0082] S140. Determine the fault type of the transformer according to the target fusion feature.

[0083] Among them, the fault type can be understood as the fault state corresponding to the transformer. In this embodiment, the fault type can include partial discharge (PD) type, low energy discharge (D1) type, high energy discharge (D2) type, medium and low temperature overheating (T1) type and high temperature overheating (T2) type. In this embodiment, the target fusion features can be classified by a classification function, and the corresponding fault type is determined according to the result of the classification process.

[0084] The technical solution of the embodiment of the present invention collects a set of dissolved gas parameters in the oil of the transformer; inputs the set of dissolved gas parameters in the oil into a generative adversarial network to obtain target gas parameter features; inputs the target gas parameter features into a feature decision fusion model to obtain target fusion features; and determines the fault type of the transformer according to the target fusion features. This technical solution processes the set of dissolved gas parameters in the oil through a generative adversarial network model and a feature decision fusion model respectively, and then determines the fault type of the transformer, thereby improving the accuracy and reliability of transformer fault detection.

[0085] Embodiment 2

[0086] Figure 3 This is a flow chart of a transformer fault detection method provided according to the second embodiment of the present invention. This embodiment is optimized based on the above embodiment. The specific optimization is as follows: the feature decision fusion model includes a feature fusion layer; the target gas parameter feature is input into the feature decision fusion model to obtain the target fusion feature, including: performing feature fusion processing on the target gas parameter feature through the feature fusion layer to obtain initial fusion feature data; channel connection between the initial fusion feature data and the target gas parameter feature to obtain the target fusion feature. Figure 3 As shown, the method includes:

[0087] S310. Collect a set of parameters of dissolved gases in transformer oil.

[0088] S320, inputting the parameter set of dissolved gas in oil into a generative adversarial network to obtain target gas parameter characteristics.

[0089] In this embodiment, the generative adversarial network is constructed based on the generative adversarial network structure. Specifically, the generative adversarial network may include a generator G, a discriminator D, and a reconstruction network R. In the training process of the generative adversarial network, the generator G and the discriminator D are optimized by minimizing the following loss function:

[0090]

[0091] Among them, L adv represents the adversarial loss, p z (z) is the distribution of the latent variable z, D is the discriminator, G is the generator, and x is the real sample;

[0092]

[0093] Among them, L rec represents the reconstruction loss, R is the reconstruction network, and ||·||2 represents the L2 norm;

[0094]

[0095] Among them, L msff represents the multi-scale feature fusion loss, N iPj and N iTj denote the number of remaining samples of the i-th category in the j-th experiment generated by the proposed method and the traditional method, respectively, C is the number of categories, and M is the number of experiments;

[0096] L total =L adv +λ rec L rec +λ msff L msff ;

[0097] Among them, L total represents the total loss, λ rec and λ msff are the weight coefficients of reconstruction loss and multi-scale feature fusion loss respectively; by optimizing the above loss function, the generator G is trained to generate high-quality fault samples, which are superior to those generated by traditional GAN ​​models in terms of similarity and diversity.

[0098] Furthermore, in this embodiment, regularization technology is introduced to optimize the generative adversarial network to prevent the problem of mode collapse. This regularization technology is intended to encourage the generator to generate more diverse samples, thereby avoiding the problem of only learning a few modes in the data set. The regularization term is achieved by maximizing the distance between generated samples. This regularization term can be defined as:

[0099]

[0100] Among them, z and z′ are from p z (z), d(·,·) is a metric function used to calculate the distance between two generated samples G(z) and G(z′). The metric function d(·,·) is selected as the Euclidean distance metric.

[0101] Therefore, the total loss function of the generator can be expressed as:

[0102] L G =L adv +λ rec L rec +λ msff L msff +λ div R diversity ;

[0103] Among them, L adv is the loss against which L rec is the reconstruction loss, L msff is the multi-scale feature matching loss, and R diversity This is the regularization term we added. Each loss term has its corresponding weight coefficient λ rec ,λ msff ,λ div , these weights determine the impact of different loss terms on the final result.

[0104] In this embodiment, the generative adversarial network minimizes L G , the generator G is trained to generate high-quality and diverse samples that can better cover all modes in the dataset, thus avoiding the problem of mode collapse.

[0105] In this embodiment, the feature decision fusion model includes a feature fusion layer.

[0106] S330, performing feature fusion processing on the target gas parameter features through the feature fusion layer to obtain initial fusion feature data.

[0107] Among them, the feature fusion layer can be used to perform feature fusion operations on the input target gas parameter features. The feature fusion processing can be a processing operation of rescaling and fusing the target gas parameter features. In this embodiment, the target gas parameter features may include gas parameter features of different gas change characteristics, and the gas parameter features of different gas change characteristics are subjected to feature fusion processing through the feature fusion layer to obtain initial fusion feature data.

[0108] In this embodiment, optionally, feature fusion processing is performed on the target gas parameter characteristics through a feature fusion layer to obtain initial fused feature data, including: determining the gas change characteristics of the target gas parameter characteristics; determining the first-level gas characteristics and the second-level gas characteristics according to the gas change characteristics; and feature fusing the first-level gas characteristics and the second-level gas characteristics to obtain initial fused feature data.

[0109] Among them, the gas change characteristics may include fast change characteristics and slow change characteristics. In this embodiment, the fast change characteristics and slow change characteristics corresponding to the target gas parameter characteristics can be received respectively through different receiving domains to determine different gas change characteristics. The first-level gas characteristics may refer to shallow-level characteristics, which are used to capture the fast change characteristics of the gas concentration in the target gas parameter characteristics. The second-level gas characteristics may refer to deep-level characteristics, which are used to capture the additional slow change characteristics of the gas concentration in the target gas parameter characteristics. Specifically, in this embodiment, the shallow characteristics correspond to a small receiving domain, which is used to extract the fast change characteristics of the gas concentration in the target gas parameter characteristics; the deep characteristics have a larger receiving domain, which is used to capture the additional slow change characteristics of the gas concentration in the target gas parameter characteristics to determine the first-level gas characteristics and the second-level gas characteristics. In this embodiment, the specific method of feature fusion of the first-level gas characteristics and the second-level gas characteristics can be to adjust the scale of the first-level gas characteristics so that the adjusted first-level gas characteristics and the second-level gas characteristics are feature-connected, thereby realizing the feature fusion processing and obtaining the initial fusion feature data.

[0110] Through such a setting in this embodiment, gas features of different levels can be captured according to the gas change characteristics, and then initial fused feature data can be obtained through feature fusion, thereby further improving the feature classification capability and enhancing the accuracy of the feature data.

[0111] In this embodiment, optionally, the first-level gas features and the second-level gas features are feature fused to obtain initial fused feature data, including: performing a scale reconstruction operation on the first-level gas features to obtain reconstructed gas features; and fusing the reconstructed gas features with the second-level gas features to obtain initial fused feature data.

[0112] Among them, the scale reconstruction operation can be understood as an operation of adjusting the scale of the first-level gas feature. In this embodiment, the specific method of performing the scale reconstruction operation on the first-level gas feature can be to perform a convolution operation on the first-level gas feature to reconstruct the feature scale to obtain the reconstructed gas feature. In this embodiment, the specific method of fusing the reconstructed gas feature with the second-level gas feature can be to connect the reconstructed gas feature with the second-level gas feature through a channel, and then further convolve to complete feature fusion, thereby obtaining initial fused feature data.

[0113] In this embodiment, the first-level gas features are processed by a lower-level feature layer, and the second-level gas features are processed by a higher-level feature layer. Specifically, in this embodiment, the feature scale is reconstructed by performing a convolution operation on the first-level gas features at the lower-level feature layer to obtain the reconstructed gas features, and the reconstructed gas features are promoted to be connected with the second-level gas features processed by the higher-level feature layer to obtain the initial fused feature data.

[0114] In this embodiment, the feature decision fusion model is used to input the output of the generative adversarial network into the decision layer of the feature decision fusion model to integrate features from different scales without significantly increasing the computational cost and information loss. In addition, the feature decision fusion model in this embodiment replaces the convolutional layer in the traditional discriminator network to significantly increase the number of feature layers and improve the classification ability of features; the performance of the deep network is improved by fusing features of different scales.

[0115] S340, performing channel connection on the initial fusion feature data and the target gas parameter feature to obtain the target fusion feature.

[0116] The target fusion feature can be understood as feature data obtained by connecting the initial fusion feature and the target gas parameter feature along the channel. In this embodiment, the initial fusion feature data and the target gas parameter feature can be connected along the channel axis, and then a convolution operation or the like is performed to obtain the target fusion feature.

[0117] Exemplarily, in this embodiment, the reconstructed gas feature is promoted to be connected with the second-level gas feature processed by the high-level feature layer to obtain the initial fusion feature data, as follows:

[0118] Di′=ReLU(BN(W*Di));

[0119] Then the initial fusion feature data and the target gas parameter feature are connected along the channel axis to obtain the target fusion feature:

[0120] Di=Concat(F,D,axis=Channel);

[0121] Among them, Concat represents the connection along the channel axis, BN represents batch normalization, ReLU represents the activation function, and W represents the weight of the convolution operation.

[0122] S350. Determine the fault type of the transformer according to the target fusion feature.

[0123] In this embodiment, optionally, determining the fault type of the transformer according to the target fusion feature includes: processing the target fusion feature through a classification function to obtain the fault type of the corresponding transformer.

[0124] The classification function processing may be an activation classification function or other classification functions. Exemplarily, in this embodiment, the classification function may be a ReLU activation function.

[0125] In this embodiment, after the target fusion features are classified through the ReLU activation function, each target fusion feature will obtain 5 neurons, each neuron corresponds to a fault type, and the weight corresponding to each neuron is different. The fault type represented by the neuron with the largest weight is determined as the fault type corresponding to the transformer.

[0126] Through such a setting, this embodiment can classify and determine the transformer fault type according to the target fusion features, thereby improving the accuracy of transformer fault detection.

[0127] Furthermore, in this embodiment, the generative adversarial network and the feature fusion model can be combined into an overall data processing model. In this embodiment, during the training of the overall model, a reconstruction network is established by minimizing the similarity between the encoding vector and the latent vector of the generated gas parameter sample to suppress the error of the feature vector reconstructed by the generated gas parameter sample. Among them, the encoding vector refers to the representation obtained after the generated sample is encoded. The encoding process is usually a process of mapping data from a high-dimensional space to a low-dimensional space, with the purpose of extracting key features of the data. The latent vector is a random noise vector used in the generative model. Reconstruct the network, which receives the generated samples and attempts to restore the initial random noise (i.e., the latent vector) used to generate these samples, thereby reducing the difference between the encoded representation of the generated samples and these latent vectors. This method helps to ensure the diversity of the generated samples while ensuring that they are sufficiently similar to the actual samples.

[0128] In this embodiment, the reconstruction network includes establishing a reconstruction network R to limit the error of the latent vector z reconstructed by the generated sample to prevent overfitting of the generated sample and improve its diversity; by minimizing the Euclidean distance between the latent vector and the original latent vector:

[0129]

[0130] In addition, in order to improve the stability and convergence speed of the model, spectral normalization is introduced to impose relatively mild constraints on the model parameters; different learning rate update rules are used to improve the convergence performance of the model.

[0131] The technical solution of this embodiment uses a multi-parameter coupled oil-immersed transformer intelligent sensor device based on a micro-electromechanical system to collect a sequence of dissolved gas parameters in oil. Subsequently, a generative adversarial network structure integrating feature extraction and feature fusion is constructed to extract features from the sequence of dissolved gas parameters in oil. In order to further improve the diagnostic accuracy, a feature fusion model of multi-scale feature decision fusion is designed. The lower-level feature layer is processed by convolution operation, and the scale of the feature is adjusted by using these processed features as weights, and then connected with the higher-level features and feature fusion is completed to determine the fault type of the transformer. In addition, a reconstruction network is established by minimizing the similarity between the encoding vector and the latent vector of the generated gas sample to suppress the error of the feature vector reconstructed by the generated sample. In the model training process in this embodiment, the dissolved gas data sets in the power transformer oil under different environmental working conditions are selected respectively, which are data under various fault types of the transformer, specifically including sample data under fault types such as partial discharge (PD), low-energy discharge (D1), high-energy discharge (D2), medium and low temperature overheating (T1) and high temperature overheating (T2). Each fault mode contains 8192 data points and is organized into 100 samples. Therefore, this embodiment has significant advantages in terms of similarity, diversity and effectiveness of generated samples, and can therefore be effectively applied to transformer fault diagnosis, greatly improving the accuracy of diagnosis.

[0132] The technical solution of the embodiment of the present invention collects a set of dissolved gas parameters in the oil of the transformer; inputs the set of dissolved gas parameters in the oil into a generative adversarial network to obtain target gas parameter features; performs feature fusion processing on the target gas parameter features through a feature fusion layer to obtain initial fusion feature data; performs channel connection between the initial fusion feature data and the target gas parameter features to obtain target fusion features; and determines the fault type of the transformer according to the target fusion features. This technical solution processes the set of dissolved gas parameters in the oil through a generative adversarial network model and a feature decision fusion model respectively, and then determines the fault type of the transformer, thereby improving the accuracy and reliability of transformer fault detection.

[0133] Embodiment 3

[0134] Figure 4 1 is a schematic diagram of the structure of a transformer fault detection device provided according to Embodiment 3 of the present invention. Figure 4 As shown, the device comprises:

[0135] The data acquisition module 410 is used to collect a set of parameters of dissolved gases in transformer oil;

[0136] The gas parameter feature acquisition module 420 is used to input the set of dissolved gas parameters in oil into the generative adversarial network to obtain the target gas parameter features;

[0137] The target fusion feature acquisition module 430 is used to input the target gas parameter feature into the feature decision fusion model to obtain the target fusion feature;

[0138] The fault type determination module 440 is used to determine the fault type of the transformer according to the target fusion feature.

[0139] Optionally, the generative adversarial network includes multiple convolution kernels; the gas parameter feature acquisition module 320 includes:

[0140] A convolution processing unit, used for performing convolution processing on a set of dissolved gas parameters in oil through multiple convolution kernels to obtain a multi-level feature vector;

[0141] The feature vector processing unit is used to perform pooling and normalization processing on the multi-level feature vectors to obtain the target gas parameter characteristics.

[0142] Optionally, the multiple convolution kernels include a first convolution kernel, a second convolution kernel and a third convolution kernel;

[0143] The convolution processing unit is specifically used to perform convolution processing on the set of dissolved gas parameters in oil through a first convolution kernel to obtain a first eigenvector; perform convolution processing on the first eigenvector through a second convolution kernel to obtain a second eigenvector; perform convolution processing on the second eigenvector through a third convolution kernel to obtain a third eigenvector; and connect the first eigenvector, the second eigenvector and the third eigenvector through channels to obtain a multi-level eigenvector.

[0144] Optionally, the feature decision fusion model includes a feature fusion layer;

[0145] The target fusion feature acquisition module 430 includes:

[0146] A fusion processing unit, used for performing feature fusion processing on the target gas parameter features through a feature fusion layer to obtain initial fusion feature data;

[0147] The feature connection unit is used to perform channel connection on the initial fusion feature data and the target gas parameter feature to obtain the target fusion feature.

[0148] Optionally, a fusion processing unit, including:

[0149] A gas variation characteristic determination subunit, used to determine the gas variation characteristic of the target gas parameter feature;

[0150] A gas characteristic determination subunit, used to determine the first level gas characteristic and the second level gas characteristic according to the gas variation characteristics;

[0151] The gas feature fusion subunit is used to fuse the first-level gas features and the second-level gas features to obtain initial fused feature data.

[0152] Optionally, the gas feature fusion subunit is specifically used to perform a scale reconstruction operation on the first-level gas features to obtain reconstructed gas features; and fuse the reconstructed gas features with the second-level gas features to obtain initial fused feature data.

[0153] Optionally, the fault type determination module 440 is specifically configured to process the target fusion feature through a classification function to obtain the fault type of the corresponding transformer.

[0154] A transformer fault detection device provided in an embodiment of the present invention can execute a transformer fault detection method provided in any embodiment of the present invention, and has functional modules and beneficial effects corresponding to the execution method.

[0155] Embodiment 4

[0156] Figure 5 1 is a schematic diagram of the structure of an electronic device provided according to Embodiment 4 of the present invention. The electronic device 10 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0157] like Figure 5 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, ROM 12 and RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0158] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0159] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a transformer fault detection method.

[0160] In some embodiments, the transformer fault detection method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the transformer fault detection method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the transformer fault detection method in any other appropriate manner (e.g., by means of firmware).

[0161] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0162] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0163] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0164] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0165] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0166] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0167] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0168] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A transformer fault detection method, characterized in that: include: Collect the parameter set of dissolved gas in transformer oil; Inputting the set of dissolved gas parameters in the oil into a generative adversarial network to obtain target gas parameter characteristics; Inputting the target gas parameter characteristics into a feature decision fusion model to obtain a target fusion characteristic; The fault type of the transformer is determined according to the target fusion feature.

2. The method according to claim 1, characterized in that The generative adversarial network includes multiple convolution kernels; The set of dissolved gas parameters in oil is input into the generative adversarial network to obtain the target gas parameter features, including: Convolution processing is performed on the set of dissolved gas parameters in the oil by using multiple convolution kernels to obtain a multi-level feature vector; The multi-level feature vectors are pooled and normalized to obtain target gas parameter features.

3. The method according to claim 2, characterized in that The multiple convolution kernels include a first convolution kernel, a second convolution kernel and a third convolution kernel; The set of dissolved gas parameters in oil is convolved by multiple convolution kernels to obtain a multi-level feature vector, including: Performing convolution processing on the set of dissolved gas parameters in oil by using the first convolution kernel to obtain a first eigenvector; Performing convolution processing on the first eigenvector using the second convolution kernel to obtain a second eigenvector; Performing convolution processing on the second eigenvector through the third convolution kernel to obtain a third eigenvector; The first feature vector, the second feature vector and the third feature vector are channel-connected to obtain a multi-level feature vector.

4. The method according to claim 1, characterized in that: The feature decision fusion model includes a feature fusion layer; The target gas parameter feature is input into the feature decision fusion model to obtain the target fusion feature, including: Performing feature fusion processing on the target gas parameter features through the feature fusion layer to obtain initial fusion feature data; The initial fusion feature data and the target gas parameter feature are channel-connected to obtain a target fusion feature.

5. The method according to claim 4, characterized in that The feature fusion layer performs feature fusion processing on the target gas parameter features to obtain initial fusion feature data, including: Determining gas variation characteristics of the target gas parameter characteristics; Determining first-level gas characteristics and second-level gas characteristics according to the gas change characteristics; The first-level gas features and the second-level gas features are feature fused to obtain initial fused feature data.

6. The method according to claim 5, characterized in that The first-level gas feature and the second-level gas feature are subjected to feature fusion to obtain initial fusion feature data, including: Performing a scale reconstruction operation on the first-level gas features to obtain reconstructed gas features; The reconstructed gas feature is fused with the second-level gas feature to obtain initial fused feature data.

7. The method according to claim 1, characterized in that Determining the fault type of the transformer according to the target fusion feature includes: The target fusion features are processed by a classification function to obtain the fault type of the corresponding transformer.

8. A transformer fault detection device, characterized in that: include: Data acquisition module, used to collect the set of dissolved gas parameters in transformer oil; A gas parameter feature acquisition module, used for inputting the set of dissolved gas parameters in the oil into a generative adversarial network to obtain target gas parameter features; A target fusion feature acquisition module is used to input the target gas parameter feature into a feature decision fusion model to obtain a target fusion feature; A fault type determination module is used to determine the fault type of the transformer according to the target fusion feature.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the transformer fault detection method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the transformer fault detection method according to any one of claims 1 to 7 when executed.

Images