A method and system for comprehensive error status assessment of capacitive voltage transformers

Abstract

This invention relates to the field of online power metering monitoring technology, and provides a method and system for comprehensive error status assessment of capacitive voltage transformers (CVTs). The method includes: acquiring the online waveform signal of the CVT, performing slicing and signal processing to obtain time-spectrum features; based on the time-spectrum features, using several convolutional neural networks to extract daily CVT features; adding temperature and humidity information to the daily CVT features according to time correspondence to construct comprehensive daily CVT features; based on the comprehensive daily CVT features, using a fully connected layer to obtain first intermediate features; based on the historical daily comprehensive CVT features, using a bidirectional long short-term memory network to obtain the output result, which is then passed through a fully connected layer to obtain second intermediate features; fusing the first and second intermediate features to obtain fused features; obtaining an evaluation result based on the fused features; and optimizing the parameters of the adaptive neural network model based on the evaluation result.

Description

Technical Field

[0001] This invention relates to the field of online power metering and monitoring technology, and in particular to a method and system for comprehensive error status assessment of capacitive voltage transformers. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] As a crucial component of electricity metering devices, the accuracy and reliability of instrument transformer (ET) metering performance directly impact the fairness and impartiality of electricity trade settlement. A capacitive voltage transformer (CVT) consists of a series capacitor that divides the voltage, which is then stepped down and isolated by an electromagnetic instrument transformer. As an instrument used for voltage transformation, the capacitive voltage transformer can also couple carrier frequencies to transmission lines for long-distance communication, selective high-frequency line protection, remote control, and other functions. Compared to conventional electromagnetic voltage transformers, capacitive voltage transformers offer advantages such as high impulse insulation strength, simple manufacturing, small size, and light weight, as well as significant economic and safety benefits.

[0004] In actual CVT operation, CVT errors are affected by factors such as the acquisition principle and working environment, leading to varying degrees of wear and tear over long-term operation. Due to the difficulty of power outages on high-voltage transmission lines, offline CVT monitoring cannot be performed frequently, making it impossible to promptly determine whether the CVT is in normal operating condition. This affects fair trade settlement of electricity and poses potential fault hazards.

[0005] Existing technologies involve training a neural network model using offline CVT data from a specific scenario. Once trained, this model is directly used to evaluate the overall error state of online CVT data. However, each CVT has a different usage scenario, so the trained neural network model cannot be universally applicable to all CVT scenarios, resulting in significant errors in the evaluation results. Summary of the Invention

[0006] To address the technical problems mentioned above, this invention provides a comprehensive error state assessment method and system for capacitive voltage transformers. By using offline CVT data from multiple scenarios to adaptively adjust the neural network model, the accuracy of the model's prediction of CVT data under various scenarios is improved.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] The first aspect of the present invention provides a method for evaluating the overall error status of a capacitive voltage transformer.

[0009] A method for comprehensive error status assessment of capacitive voltage transformers includes:

[0010] The daily real-time waveform signals, temperature information, and humidity information of the CVT are acquired, and a trained adaptive neural network model is used to obtain the comprehensive error status evaluation results of the capacitive voltage transformer.

[0011] The training process of the adaptive neural network model includes: acquiring the daily real-time waveform signals of the CVT, and performing slicing and signal processing to obtain time-spectrum features; based on the time-spectrum features, using several convolutional neural networks to extract CVT daily features; adding temperature and humidity information to the CVT daily features according to the time correspondence to construct CVT daily comprehensive features; based on the CVT daily comprehensive features, using a fully connected layer to obtain the first intermediate feature; based on the CVT historical daily comprehensive features, using a bidirectional long short-term memory network to obtain the output result, and then passing it through a fully connected layer to obtain the second intermediate feature; fusing the first and second intermediate features to obtain the fused feature; based on the fused feature, obtaining the evaluation result, and optimizing the parameters of the adaptive neural network model according to the evaluation result.

[0012] Furthermore, the slicing process includes: setting a slice length, and having a set time interval between two adjacent slices for repeated extraction, and slicing the daily real-time acquired waveform signal of the CVT.

[0013] Furthermore, the process of extracting CVT daily features based on time-frequency spectral features using several convolutional neural networks includes: obtaining several features based on time-frequency spectral features using several convolutional neural networks; and vertically superimposing the several features to obtain CVT daily features.

[0014] Furthermore, the process of adding temperature and humidity information to the daily CVT features according to the time correspondence to construct the daily CVT comprehensive features includes: calculating the average temperature and average humidity for a certain time period based on the daily temperature and humidity curves; adding the average temperature and average humidity to the daily CVT features based on the time correspondence between the certain time period and the time in the daily CVT features to construct the daily CVT comprehensive features.

[0015] Furthermore, the process of fusing the first intermediate feature and the second intermediate feature to obtain the fused feature adopts the following formula:

[0016] o i =w i1 l i +w i2 r i

[0017] Among them, o i Let l represent the i-th dimension component of the fused feature O. iLet r represent the i-th dimension component of the daily comprehensive feature L. i w represents the i-th dimension component of the calculated daily feature R based on historical operating states. i1 ,w i2 They represent the i-th dimension components of the daily comprehensive feature, respectively. i The weights and calculation of the i-th dimension component r of the daily feature based on historical operating status i The weight.

[0018] Furthermore, a dataset is constructed based on the time-spectrum features and the state categories corresponding to the time-spectrum features; the dataset is divided into a training set and a test set; and the adaptive neural network model is trained using the training set.

[0019] Furthermore, the process of optimizing the parameters of the adaptive neural network model based on the evaluation results includes determining whether the evaluation results are the same as the state category. If they are not the same, the parameters of the adaptive neural network model are optimized. If they are the same, the next round of iterative training continues until the set number of iterations is reached, and a trained adaptive neural network model is obtained.

[0020] A second aspect of the present invention provides a comprehensive error status assessment system for a capacitive voltage transformer.

[0021] A comprehensive error status assessment system for capacitive voltage transformers includes:

[0022] The online evaluation module is configured to: acquire daily real-time waveform signals, temperature information and humidity information of the CVT, and use a trained adaptive neural network model to obtain the comprehensive error status evaluation results of the capacitive voltage transformer.

[0023] The adaptive neural network model training module is configured to: acquire daily real-time waveform signals of the CVT, perform slicing and signal processing to obtain time-spectrum features; based on the time-spectrum features, use several convolutional neural networks to extract daily CVT features; add temperature and humidity information to the daily CVT features according to time correspondence to construct comprehensive daily CVT features; based on the comprehensive daily CVT features, use a fully connected layer to obtain the first intermediate feature; based on the historical daily comprehensive CVT features, use a bidirectional long short-term memory network to obtain the output result, and then pass it through a fully connected layer to obtain the second intermediate feature; fuse the first and second intermediate features to obtain the fused feature; based on the fused feature, obtain the evaluation result, and optimize the parameters of the adaptive neural network model according to the evaluation result.

[0024] A third aspect of the present invention provides a computer-readable storage medium.

[0025] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the comprehensive error state assessment method for capacitive voltage transformers as described in the first aspect above.

[0026] A fourth aspect of the present invention provides a computer device.

[0027] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the steps in the comprehensive error state assessment method for capacitive voltage transformers as described in the first aspect above.

[0028] Compared with the prior art, the beneficial effects of the present invention are:

[0029] This invention uses offline CVT waveform signals from multiple scenarios to adaptively adjust a neural network model. The resulting adaptive neural network model is used for real-time prediction of online CVT waveform signals, thereby improving the model's prediction accuracy for online CVT waveforms in various scenarios, promptly determining whether the CVT is in a normal operating state, and avoiding potential faults. Attached Figure Description

[0030] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0031] Figure 1 This is a flowchart of the comprehensive error state assessment method for capacitive voltage transformers shown in this invention;

[0032] Figure 2 This is a structural diagram of the deep learning feature extraction model shown in this invention;

[0033] Figure 3 This is a flowchart illustrating the adaptive adjustment of CVT as shown in this invention. Detailed Implementation

[0034] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0035] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0036] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0037] It should be noted that the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of methods and systems according to various embodiments of this disclosure. It should be noted that each block in a flowchart or block diagram may represent a module, segment, or portion of code, which may include one or more executable instructions for implementing the logical functions specified in the various embodiments. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that shown in the drawings. For example, two consecutively represented blocks may actually be executed substantially in parallel, or they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the flowcharts and / or block diagrams, and combinations of blocks in the flowcharts and / or block diagrams, may be implemented using a dedicated hardware-based system that performs the specified functions or operations, or using a combination of dedicated hardware and computer instructions.

[0038] Example 1

[0039] like Figure 1 As shown, this embodiment provides a comprehensive error state assessment method for capacitive voltage transformers. This embodiment uses the application of this method to a server as an example for illustration. It is understood that this method can also be applied to terminals, and can also be applied to systems including terminals, servers, and other components, and can be implemented through interaction between the terminal and the server. The server can be an independent physical server, a server cluster composed of multiple physical servers, or a distributed system. It can also be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network servers, cloud communication, middleware services, domain name services, CDN security services, and big data and artificial intelligence platforms. The terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, etc., but is not limited to these. The terminal and server can be directly or indirectly connected via wired or wireless communication, which is not limited herein. In this embodiment, the method includes the following steps:

[0040] The daily real-time waveform signals, temperature information, and humidity information of the CVT are acquired, and a trained adaptive neural network model is used to obtain the comprehensive error status evaluation results of the capacitive voltage transformer.

[0041] The training process of the adaptive neural network model includes: acquiring the daily real-time waveform signals of the CVT, and performing slicing and signal processing to obtain time-spectrum features; based on the time-spectrum features, using several convolutional neural networks to extract CVT daily features; adding temperature and humidity information to the CVT daily features according to the time correspondence to construct CVT daily comprehensive features; based on the CVT daily comprehensive features, using a fully connected layer to obtain the first intermediate feature; based on the CVT historical daily comprehensive features, using a bidirectional long short-term memory network to obtain the output result, and then passing it through a fully connected layer to obtain the second intermediate feature; fusing the first and second intermediate features to obtain the fused feature; based on the fused feature, obtaining the evaluation result, and optimizing the parameters of the adaptive neural network model according to the evaluation result.

[0042] Step 1: Data collection and preprocessing.

[0043] The daily CVT waveform signal is collected from the monitoring device of the CVT current transformer, and the corresponding daily CVT status is obtained based on the results of the CVT online monitoring platform. The daily CVT waveform signal is sliced ​​according to the following rules: the slice length is set to T(s), and there is a repetition extraction length of Δt between two adjacent slices. N waveform signal segments are generated. The time spectrum S = {s1, s2, ..., s...} of each waveform signal segment is obtained through corresponding signal processing operations. N The dataset D = {X, Y} is set up as follows: any sample d(x, y), (d ∈ D, x ∈ X, y ∈ Y) represents data within any date, and its features are represented by the daily hourly spectrum, i.e., X = S. Its classification state Y is represented by the daily state displayed in the online transformer detection system. The corresponding categories are (normal, abnormal, alarm). The constructed dataset is normalized and divided into a training set D in a 7:3 ratio. training and test set D test .

[0044] When the monitoring device collects data for a short period, resulting in insufficient data volume, duplicate data extraction between slices can generate more segments, leading to more thorough training. Waveform signal data stores time-dimension information. After various signal processing operations, non-repeating slices weaken the temporal sequence correspondence between slices, i.e., weaken the time-dimension information. Therefore, to retain relevant information, waveform data undergoes repeated segment slicing.

[0045] Step 2: Construct a deep learning feature extraction model. The constructed deep learning feature extraction model is as follows: Figure 2 As shown.

[0046] Step 2.1: Analyze the characteristic time-frequency spectrum {s1,s2,…,s} of sample data d. N Input the corresponding N convolutional neural network (CNN) modules. In each CNN module, the time-frequency spectrum s is processed through multiple convolutional pooling layers, and the deep features are expanded by fully connected layers to obtain k-dimensional features q, dim = [1×k]. The features {q1, q2, ..., q} extracted from the N modules are then processed. N The CVT daily feature f,dim=[N×k] is obtained by vertical stacking.

[0047] Step 2.2: Add temperature and humidity information. Based on the daily temperature and humidity curves, obtain the time period T. D The average temperature and humidity. T D The correspondence between temperature and humidity (T) is used to add temperature and humidity information to the daily characteristics of the CVT. The new daily characteristic f′ is then...

[0048]

[0049] in, This represents the dimensionality feature of the corresponding CNN module after processing the i-th slice. This represents the average temperature of the time period corresponding to the i-th slice. This represents the average humidity value extracted from the time period corresponding to the i-th slice.

[0050] Step 3: Construct a deep learning classification model, which is a deep learning feature classification model.

[0051] Step 3.1: Extract CVT features for day D, F = {f1′, f2′, ..., f D The result is fed into a two-layer Bi-LSTM model for classification training.

[0052] Step 3.2: By repeating the steps in Step 2 and Step 3.1, input set D. Train The feature data x and label data y of each sample are used for model training to obtain a suitable network model.

[0053] Step 3.3: Test set D Test The sample is input into the trained network model to obtain the predicted label y of the corresponding sample. predict , will y predict With actual label y test The accuracy of the proposed solution can be determined through comparison.

[0054] Step 4: Adaptively adjust the CVT.

[0055] Since each CVT has a different use case, in order to accurately evaluate the overall error status of each CVT, the model is adaptively adjusted. The process is as follows: Figure 3 As shown.

[0056] Step 4.1: Based on the existing model, the comprehensive error evaluation model of the CVT under test is adaptively adjusted according to the historical data of the CVT under test. Based on the previous model and parameters, the comprehensive error evaluation model of the CVT under test is adjusted as follows: the daily comprehensive features are calculated by passing them through a Dense layer to obtain the daily comprehensive features L={l1,l2,…,l k This is used to enhance the impact of the computation day's operation process, and the Dense layer output by the Bi-LSTM is connected to another Dense layer to obtain the computation day features R = {r1, r2, ..., r} based on historical operating states. k This is used to strengthen the influence of historical data on the calculation date during this period. The two strongly influencing components are fused to obtain O, specifically through the following method:

[0057] o i =w i1 l i +w i2 r i

[0058] Among them, w i1 w represents the weight of the i-th dimension component in the calculation of the daily comprehensive feature. i2 This represents the weight of the i-th dimension component of the calculated daily feature based on historical operating states.

[0059] That is, O = {o1, o2, ..., o} k}={w 11 l1+w 12 r1,w 21 l2+w 22 r2,…,w k1 l k +w k2 r k}

[0060] Among them, o i Let l represent the i-th dimension component of the fused feature O. i Let r represent the i-th dimension component of the daily comprehensive feature L. i w represents the i-th dimension component of the calculated daily feature R based on historical operating states. i1 ,w i2 They represent the i-th dimension components of the daily comprehensive feature, respectively. i The weights and calculation of the i-th dimension component r of the daily feature based on historical operating statusi The weight.

[0061] The final evaluation result is obtained through the softmax function.

[0062] This invention's adaptive adjustment, based on daily data collection, detects a certain degree of abrupt change in the CVT under test. To mitigate the impact of sudden changes in the secondary circuit and other factors, it incorporates general results based on the historical operating status of the CVT under test. This allows for a comprehensive assessment of the CVT's operating status. By analyzing the changes in the CVT's operating status over time, this historical analysis, compared to analyzing the CVT's operating status only on the test date, better reflects the general characteristics of the CVT under test and mitigates the impact of sudden changes in real-time data due to specific operating conditions.

[0063] Step 4.2: Following the process in Step 4.1, using the historical data of the CVT to be tested, and according to the pre-constructed overall CVT evaluation model and parameters, the evaluation model of the CVT to be tested is trained and adjusted to obtain a model with the highest accuracy.

[0064] Step 5: Use a deep neural network model to perform online comprehensive error assessment and classification of CVT.

[0065] The system collects waveform signals, temperature, and humidity information corresponding to the CVT online monitoring system for the date to be measured. After data preprocessing, the data is input into an adaptive neural network model, which outputs the classification results of the CVT online comprehensive error assessment for that day.

[0066] Example 2

[0067] This embodiment provides a comprehensive error status assessment system for capacitive voltage transformers.

[0068] A comprehensive error status assessment system for capacitive voltage transformers includes:

[0069] The online evaluation module is configured to: acquire daily real-time waveform signals, temperature information and humidity information of the CVT, and use a trained adaptive neural network model to obtain the comprehensive error status evaluation results of the capacitive voltage transformer.

[0070] The adaptive neural network model training module is configured to: acquire daily real-time waveform signals of the CVT, perform slicing and signal processing to obtain time-spectrum features; based on the time-spectrum features, use several convolutional neural networks to extract daily CVT features; add temperature and humidity information to the daily CVT features according to time correspondence to construct comprehensive daily CVT features; based on the comprehensive daily CVT features, use a fully connected layer to obtain the first intermediate feature; based on the historical daily comprehensive CVT features, use a bidirectional long short-term memory network to obtain the output result, and then pass it through a fully connected layer to obtain the second intermediate feature; fuse the first and second intermediate features to obtain the fused feature; based on the fused feature, obtain the evaluation result, and optimize the parameters of the adaptive neural network model according to the evaluation result.

[0071] It should be noted that the online evaluation module and adaptive neural network model training module described above are the same examples and application scenarios implemented in Embodiment 1, but are not limited to the content disclosed in Embodiment 1. It should also be noted that these modules, as part of the system, can be executed in a computer system such as a set of computer-executable instructions.

[0072] Example 3

[0073] This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in the comprehensive error state assessment method for capacitive voltage transformers as described in Embodiment 1 above.

[0074] Example 4

[0075] This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the comprehensive error state evaluation method for capacitive voltage transformers as described in Embodiment 1 above.

[0076] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0077] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0078] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0079] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0080] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0081] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for comprehensive error status assessment of a capacitive voltage transformer, characterized in that, include: The daily real-time waveform signals, temperature information, and humidity information of the CVT are acquired, and a trained adaptive neural network model is used to obtain the comprehensive error status evaluation results of the capacitive voltage transformer. The training process of the adaptive neural network model includes: acquiring the daily real-time acquisition waveform signal of the CVT, and performing slicing and signal processing to obtain time-spectrum features; Based on time-frequency characteristics, several convolutional neural networks are used to extract CVT daily features; the process of extracting CVT daily features includes: processing sample data... The time spectrum of the feature s is input into the corresponding N CNN modules respectively. In each CNN module, the time spectrum s is processed by multiple convolutional pooling. The deep features are expanded by the fully connected layer to obtain the k-dimensional features q. The features extracted from the N modules are vertically stacked to obtain the CVT daily features. Temperature and humidity information are added to the daily CVT features according to the time correspondence to construct a comprehensive daily CVT feature. Based on the comprehensive features of CVT daily data, a fully connected layer is used to obtain the first intermediate feature; based on the comprehensive features of CVT historical data daily data, a bidirectional long short-term memory network is used to obtain the output result, which is then passed through a fully connected layer to obtain the second intermediate feature; The first and second intermediate features are fused to obtain the fused features; based on the fused features, the evaluation results are obtained; and based on the evaluation results, the parameters of the adaptive neural network model are optimized. The process of optimizing the parameters of the adaptive neural network model based on the evaluation results includes determining whether the evaluation results are the same as the state category. If they are not the same, the parameters of the adaptive neural network model are optimized. If they are the same, the next round of iterative training continues until the set number of iterations is reached, and the trained adaptive neural network model is obtained.

2. The method for comprehensive error status assessment of capacitive voltage transformers according to claim 1, characterized in that, The slicing process includes: setting the slice length, and having a set time interval between two adjacent slices for repeated extraction, and slicing the daily real-time acquired waveform signal of the CVT.

3. The method for comprehensive error status assessment of capacitive voltage transformers according to claim 1, characterized in that, The process of adding temperature and humidity information to the daily CVT features according to the time correspondence to construct the daily CVT comprehensive features includes: calculating the average temperature and average humidity for a certain time period based on the daily temperature and humidity curves; adding the average temperature and average humidity to the daily CVT features based on the time correspondence between the certain time period and the time in the daily CVT features to construct the daily CVT comprehensive features.

4. The method for comprehensive error status assessment of capacitive voltage transformers according to claim 1, characterized in that, The process of fusing the first intermediate feature and the second intermediate feature to obtain the fused feature uses the following formula: in, Indicates fusion features The dimensional components, Indicates the calculation of daily comprehensive characteristics The dimensional components, Represents the daily characteristics calculated based on historical operating status. The dimensional components, , These represent the calculation of the daily comprehensive characteristics. Dimensional components The weights and daily characteristics calculated based on historical operating status. Dimensional components The weight.

5. The method for comprehensive error status assessment of capacitive voltage transformers according to claim 1, characterized in that, A dataset is constructed based on the time-spectrum characteristics and the corresponding state categories; the dataset is divided into a training set and a test set; and an adaptive neural network model is trained using the training set.

6. A comprehensive error status assessment system for a capacitive voltage transformer, employing the comprehensive error status assessment method for a capacitive voltage transformer as described in any one of claims 1-5, characterized in that, include: The online evaluation module is configured to: acquire daily real-time waveform signals, temperature information and humidity information of the CVT, and use a trained adaptive neural network model to obtain the comprehensive error status evaluation results of the capacitive voltage transformer. The adaptive neural network model training module is configured to: acquire the daily real-time acquisition waveform signal of the CVT, and perform slicing and signal processing to obtain time-spectrum features; based on the time-spectrum features, use several convolutional neural networks to extract the daily features of the CVT; and add temperature and humidity information to the daily features of the CVT according to the time correspondence to construct the comprehensive daily features of the CVT. Based on the comprehensive features of CVT daily data, a fully connected layer is used to obtain the first intermediate feature; Based on the comprehensive features of CVT historical single day, a bidirectional long short-term memory network is used to obtain the output results, which are then passed through a fully connected layer to obtain the second intermediate features; The first intermediate feature and the second intermediate feature are fused to obtain the fused feature; Based on the fusion features, the evaluation results are obtained, and the parameters of the adaptive neural network model are optimized according to the evaluation results.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the comprehensive error state evaluation method for capacitive voltage transformers as described in any one of claims 1-5.

8. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the comprehensive error state evaluation method for capacitive voltage transformers as described in any one of claims 1-5.

Images