Current transformer error state evaluation method and system based on multi-stage CNN cascade

Abstract

The invention belongs to the technical field of power monitoring, and discloses a current transformer error state evaluation method and system based on multistage CNN cascading, and the method comprises the steps: collecting the historical operation current time sequence data of a current transformer, and carrying out the processing of the time sequence data, and obtaining a time domain image and a frequency domain image; performing feature extraction on the time domain image and the frequency domain image by using a multi-stage cascade CNN algorithm to obtain a time domain feature and a frequency domain feature, and fusing the time domain feature and the frequency domain feature to obtain a mixed feature; and inputting the time domain features and the frequency domain features processed by the first-stage CNN and the mixed features processed by the multi-stage CNN into an improved LSTM model to obtain an error state of the current transformer. According to the method, the time domain information and the frequency domain information of the secondary output data of the current transformer are fused, and high-accuracy error state evaluation is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power monitoring, and in particular to a current transformer error state evaluation method and system based on multi-stage CNN cascade. Background Art

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

[0003] In the power system, the current transformer is a key device in the electric energy metering system. The accuracy of its metering performance is directly related to the fairness and reliability of electric energy trading. The current transformer (CT), a precision device based on the principle of electromagnetic induction, can convert a high-amplitude primary current into an easily measurable secondary current. Its main structure includes a closed iron core and a coil wound around it. However, in the actual operating environment, the metering error of CT may gradually deviate from the standard range over time due to the limitations of the acquisition mechanism and the influence of environmental conditions. This requires not only that we have accurate and rapid fault diagnosis capabilities, but also that we must always be able to grasp the real-time status of CT metering errors, so as to guide operation and maintenance personnel to reasonably plan maintenance strategies. If the abnormality of the transformer status is not detected in time, it will pose a potential threat to the overall operation of the power grid.

[0004] Therefore, ensuring the real-time accuracy of current transformers, reducing power metering losses, and ensuring the stable operation of measurement, control and protection equipment have become urgent technical challenges. How to achieve real-time online evaluation of CT error status is undoubtedly a key problem that the power industry needs to overcome. The existing regular CT error inspection has high costs, low efficiency, and poor timeliness. It is unable to detect faulty equipment in a timely manner, which seriously affects the real-time accuracy of power trade settlement. It is urgent to break through the bottleneck of online monitoring technology and realize real-time online inspection of CT error status. Summary of the invention

[0005] In order to solve the above problems, the present invention proposes a current transformer error state evaluation method and system based on multi-stage CNN cascade, which fuses the time domain information and frequency domain information of the secondary output data of the current transformer to achieve high-accuracy error state evaluation.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: In a first aspect, the present invention provides a current transformer error state evaluation method based on a multi-stage CNN cascade, comprising the following steps: Collect the current time series data of the historical operation of the current transformer, process the time series data, and obtain the time domain image and frequency domain image; A multi-level cascade CNN algorithm is used to extract features from time domain images and frequency domain images respectively to obtain time domain features and frequency domain features, which are then fused to obtain mixed features. The time domain features and frequency domain features processed by the first-level CNN and the mixed features processed by multi-level CNN are input into the improved LSTM model to obtain the error state of the current transformer.

[0007] As an optional implementation manner, the current time series data of the historical operation of the current transformer includes zero-sequence unbalanced time series data and negative-sequence unbalanced time series data.

[0008] As an optional implementation, the time series data is processed to obtain a time domain image and a frequency domain image, specifically: The Gram angle field is used to convert the time domain signal into a time domain image; Fast Fourier transform is used to convert the time domain signal into frequency domain signal, and Gram angle is used to further convert the frequency domain signal into frequency domain image.

[0009] As an optional implementation, the time domain features and the frequency domain features are fused in a cross-fusion manner.

[0010] As an optional implementation, up-sampling the time domain features and the frequency domain features is performed to obtain sampled features; Average pooling and maximum pooling operations are performed on the time domain features and frequency domain features respectively, and the obtained features are subjected to feature splicing operations. Then, after passing through a multi-layer perceptron and an activation function, the time domain feature weights and the frequency domain feature weights are generated. The sampled features are weighted using the generated feature weights to obtain the first time domain weighted features and the first frequency domain weighted features. The first time domain weighted features and the first frequency domain weighted features are fused to obtain mixed features.

[0011] As an optional implementation, the scale factor of the improved LSTM model is an adaptive scale factor, specifically:

[0012] in, and are the maximum and minimum step sizes determined empirically, , , They are zero-sequence unbalance, negative-sequence unbalance, and the reference value of the sampling current. , , are the parameters that determine the step size adjustment. , , is a hyperparameter, is the variance.

[0013] In a second aspect, the present invention provides a current transformer error state evaluation system based on a multi-stage CNN cascade, comprising: The data acquisition and preprocessing module is configured to: acquire the current time series data of the historical operation of the current transformer, process the time series data, and obtain a time domain image and a frequency domain image; The multi-level cascade module is configured to: use a multi-level cascade CNN algorithm to extract features from the time domain image and the frequency domain image respectively to obtain time domain features and frequency domain features, and fuse the time domain features and the frequency domain features to obtain mixed features; The output module is configured to: input the time domain features and frequency domain features processed by the first-level CNN and the mixed features processed by the multi-level CNN into the improved LSTM model to obtain the error state of the current transformer.

[0014] In a third aspect, the present invention provides an electronic device, including a memory and a processor, and computer instructions stored in the memory and executed on the processor, wherein when the computer instructions are executed by the processor, the following steps are performed: Collect the current time series data of the historical operation of the current transformer, process the time series data, and obtain the time domain image and frequency domain image; A multi-level cascade CNN algorithm is used to extract features from time domain images and frequency domain images respectively to obtain time domain features and frequency domain features, which are then fused to obtain mixed features. The time domain features and frequency domain features processed by the first-level CNN and the mixed features processed by multi-level CNN are input into the improved LSTM model to obtain the error state of the current transformer.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium for storing computer instructions, wherein when the computer instructions are executed by a processor, the following steps are implemented: Collect the current time series data of the historical operation of the current transformer, process the time series data, and obtain the time domain image and frequency domain image; A multi-level cascade CNN algorithm is used to extract features from time domain images and frequency domain images respectively to obtain time domain features and frequency domain features, which are then fused to obtain mixed features. The time domain features and frequency domain features processed by the first-level CNN and the mixed features processed by multi-level CNN are input into the improved LSTM model to obtain the error state of the current transformer.

[0016] In a fifth aspect, the present invention provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the following steps are implemented: Collect the current time series data of the historical operation of the current transformer, process the time series data, and obtain the time domain image and frequency domain image; A multi-level cascade CNN algorithm is used to extract features from time domain images and frequency domain images respectively to obtain time domain features and frequency domain features, which are then fused to obtain mixed features. The time domain features and frequency domain features processed by the first-level CNN and the mixed features processed by multi-level CNN are input into the improved LSTM model to obtain the error state of the current transformer.

[0017] Compared with the prior art, the present invention has the following beneficial effects: The present invention proposes a current transformer error state evaluation method and system based on multi-stage CNN cascade, designs an effective hybrid architecture, realizes mutual complementation in the feature extraction process through multiple CNN branches, and realizes two-way information exchange of time domain features and frequency domain features through the fusion of intermediate features. This fusion strategy improves the complementarity of features and enhances the model evaluation accuracy.

[0018] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0020] Figure 1 A framework diagram of a current transformer error state evaluation method based on multi-stage CNN cascade provided in Embodiment 1 of the present invention; Figure 2 It is a framework diagram of cross-fusion of time-frequency domain feature maps of the present invention; Figure 3 It is a structural schematic diagram of the improved LSTM model of the present invention. DETAILED DESCRIPTION

[0021] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0022] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0023] It should be noted that the terms used herein are only for describing specific embodiments, and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising 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.

[0024] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0025] Terminology explanation: Gramian Angular Field (GAF): is an effective method to convert one-dimensional time series data into two-dimensional image representation. Its core principle is to calculate the cosine values ​​of the angles between data points in the time series and map these values ​​to the pixels of the two-dimensional image, thereby generating an image that can reflect the dynamic and periodic characteristics of the time series.

[0026] Example 1 like Figure 1 As shown, this embodiment provides a current transformer error state evaluation method based on multi-stage CNN cascade, comprising the following steps: S1. Collecting the current time series data of the historical operation of the current transformer, processing the time series data, and obtaining a time domain image and a frequency domain image; S2, using a multi-level cascade CNN algorithm to extract features from the time domain image and the frequency domain image respectively, obtaining time domain features and frequency domain features, and fusing the time domain features and the frequency domain features to obtain mixed features; S3. The time domain features and frequency domain features processed by the first-level CNN and the mixed features processed by the multi-level CNN are input into the improved LSTM model to obtain the error state of the current transformer.

[0027] Among them, S1, collects the current time series data of the historical operation of the current transformer, processes the time series data, and obtains the time domain image and frequency domain image. Specifically: First, collect the secondary current data of the current transformer's historical operation. The secondary current data of the current transformer includes the three-phase current of A, B, and C. , , .

[0028] Collect the zero-sequence unbalanced timing data of the current transformer: , negative sequence imbalance timing data: , T represents the historical time period, , Indicates zero-sequence imbalance and negative-sequence imbalance at time t.

[0029] Zero sequence unbalance and negative sequence unbalance calculation: (1) Zero sequence unbalance: (2) Negative sequence imbalance: (3) in , , is the three-phase current, , , That is the corresponding a-phase zero-sequence component. , Respectively represent zero-sequence imbalance and negative-sequence imbalance without time dimension. , , with a time dimension T, , They represent the zero-sequence imbalance and negative-sequence imbalance at the tth moment respectively. , They represent the zero-sequence imbalance and negative-sequence imbalance in the historical time period T respectively.

[0030] The Gramian Angular Field (GAF) is used to transform the time domain signal , Convert to time domain image , , specifically: (4) (5) The time domain signal is transformed using the Fast Fourier Transform (FFT) , The frequency domain signal is converted into a frequency domain signal and further converted into a frequency domain signal using the Gram angle. , Converted frequency domain image , , specifically: (6) (7) (8) (9) S2. Use a multi-level cascade CNN algorithm to extract features from the time domain image and the frequency domain image respectively, obtain time domain features and frequency domain features, fuse the time domain features and frequency domain features, and obtain mixed features, specifically: The time domain image TG and the frequency domain image FG are recorded as: (10) (11) Feature extraction of time domain images: (12) (13) (14) Feature extraction of frequency domain images: (15) (16) Where i represents the number of cascades, represents the i-th level time domain mixed feature, represents the i-1th level time domain image feature, represents the i-1th level time-frequency mixed feature, represents the 0th level temporal image feature, represents the 0th level time-frequency mixed feature, represents the i-th level frequency domain image feature, represents the i-1th level frequency domain image feature, Represents the 0th level frequency domain image features.

[0031] The time-frequency domain features are fused by cross fusion. The specific process is as follows: Figure 2 shown.

[0032] (1) First, the time domain image features and frequency domain image features are upsampled, and then the sampled data is spliced ​​to achieve the interaction of features between the two different domains and integrate feature information of different scales and levels, which helps the model to understand the input data more comprehensively, thereby improving its performance and accuracy.

[0033] (17) (18) (19) Among them, Conv is the convolution operation, upsample is the upsampling operation, , is the time domain image feature and frequency domain image feature after sampling, || is the feature concatenation symbol; is the comparison function, Represents the weight of the i-th level fusion.

[0034] (2) Taking the time domain image features and the frequency domain image features as input, performing feature concatenation after average pooling (AP) and maximum pooling (MP), and generating time domain feature weights and frequency domain feature weights after passing through a multi-layer perceptron (MLP) and an activation function (Sigmoid); then using the generated features to weight the sampled features to obtain the first time domain weighted features and the first frequency domain weighted features; (20) (twenty one) (twenty two) (twenty three) in, represents the activation function, is the average pooling function, is the maximum pooling function, MLP is the implicit function of the multi-layer perceptron, , They are the weights of the time domain image features and the frequency domain image features respectively.

[0035] (3) Fusing the first time-domain weighted features and the first frequency-domain weighted features to obtain fused features.

[0036] (twenty four) in, is the i-th level fusion feature, is the cross product budget, It is a direct and operation.

[0037] By assigning different weights to different features, feature weighted fusion can combine the advantages of multiple features and improve the model's expressiveness. This fusion method enables the model to better capture key information in the data, thereby improving classification accuracy.

[0038] S3, input the time domain features and frequency domain features processed by the first-level CNN together with the mixed features processed by the multi-level CNN into the improved LSTM model (such as Figure 3 As shown in Figure 2, the error state of the current transformer is obtained. The specific calculation process is: The evaluation process of the LSTM used in the present invention is as follows: the sequence value of the feature sequence X at time t When input into the network, the forget gate, input gate, output gate and memory unit will be updated and calculated respectively. The specific process is as follows: Forget Gate: (25) Input Gate: (26) Temporary Memory: (27) Memory Update: (28) Output Gate: (29) Hidden state: (30) in, , , They are the information input at time t, the hidden state output, and the hidden state input at time t. , , Corresponding to different gating structures, For additional information, To memorize information, is the weight matrix, b is the bias, , The activation function unifies the data into the range of [0,1] and [-1,1] respectively.

[0039] In order to strengthen the information interaction between the current moment feature and the previous moment error state and improve the prediction accuracy, the present invention processes it through the following formula: (31) (32) in, is the activation function, is the scale factor, , are the input and output convolutional layer weights, respectively. , Bias for the input and output layers.

[0040] Since the input features and output states will change over time and show a certain trend of change, the is an adaptive scale, specifically: (33) in, and are the maximum and minimum step sizes determined empirically, , , They are the parameters that determine the step size adjustment and describe the zero-sequence imbalance, negative-sequence imbalance, and the reference value of the sampling current during the modeling period. , , It is determined only by the variance between two consecutive sampling points. , , is a hyperparameter, p=A, B, C phase, exp is an exponential function with e as the base, is the variance.

[0041] The present invention evaluates the error state of the current transformer, finds the out-of-tolerance current transformer in time, and thus ensures the stability of the power grid operation.

[0042] Define the current transformer operating state out of tolerance as a positive example, and the current transformer state normal as a negative example. The number of correct and incorrect positive and negative examples are counted respectively, and the accuracy ACC and precision P are used as the effectiveness measurement indicators. The accuracy describes the probability that the error estimation state of this method is consistent with the actual state, and the precision describes the probability that the actual current transformer is out of tolerance among all the current transformers judged to be out of tolerance. The calculation formulas are shown in equations (34) and (35): (34)

[0043] In the formula, TP: the sample is positive and the prediction result is positive; FP: the sample is negative and the prediction result is positive; TN: the sample is negative and the prediction result is negative; FN: the sample is positive and the prediction result is negative.

[0044] Ten current transformers were selected and evaluated using the method of the present invention. The accuracy ACC reached over 98.817% and the precision C reached over 99.223%.

[0045] Example 2 This embodiment provides a current transformer error state evaluation system based on a multi-stage CNN cascade, including: The data acquisition and preprocessing module is configured to: acquire the current time series data of the historical operation of the current transformer, process the time series data, and obtain a time domain image and a frequency domain image; The multi-level cascade module is configured to: use a multi-level cascade CNN algorithm to extract features from the time domain image and the frequency domain image respectively to obtain time domain features and frequency domain features, and fuse the time domain features and the frequency domain features to obtain mixed features; The output module is configured to: input the time domain features and frequency domain features processed by the first-level CNN and the mixed features processed by the multi-level CNN into the improved LSTM model to obtain the error state of the current transformer.

[0046] It should be noted that the above modules correspond to the steps described in Example 1, and the examples and application scenarios implemented by the above modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above Example 1. It should be noted that the above modules, as part of the system, can be executed in a computer system such as a set of computer executable instructions.

[0047] In further embodiments, there is also provided: An electronic device includes a memory and a processor, and computer instructions stored in the memory and executed on the processor, wherein when the computer instructions are executed by the processor, the following steps are completed: Collect the current time series data of the historical operation of the current transformer, process the time series data, and obtain the time domain image and frequency domain image; A multi-level cascade CNN algorithm is used to extract features from time domain images and frequency domain images respectively to obtain time domain features and frequency domain features, which are then fused to obtain mixed features. The time domain features and frequency domain features processed by the first-level CNN and the mixed features processed by multi-level CNN are input into the improved LSTM model to obtain the error state of the current transformer.

[0048] It should be understood that in this embodiment, the processor may be a central processing unit CPU, and the processor may also be other general-purpose processors, digital signal processors DSP, application-specific integrated circuits ASIC, off-the-shelf programmable gate arrays FPGA or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0049] The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0050] A computer-readable storage medium is used to store computer instructions, and when the computer instructions are executed by a processor, the following steps are implemented: Collect the current time series data of the historical operation of the current transformer, process the time series data, and obtain the time domain image and frequency domain image; A multi-level cascade CNN algorithm is used to extract features from time domain images and frequency domain images respectively to obtain time domain features and frequency domain features, which are then fused to obtain mixed features. The time domain features and frequency domain features processed by the first-level CNN and the mixed features processed by multi-level CNN are input into the improved LSTM model to obtain the error state of the current transformer.

[0051] The method in Example 1 can be directly embodied as a hardware processor, or a combination of hardware and software modules in the processor. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware. To avoid repetition, it is not described in detail here.

[0052] A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the following steps are implemented: Collect the current time series data of the historical operation of the current transformer, process the time series data, and obtain the time domain image and frequency domain image; A multi-level cascade CNN algorithm is used to extract features from time domain images and frequency domain images respectively to obtain time domain features and frequency domain features, which are then fused to obtain mixed features. The time domain features and frequency domain features processed by the first-level CNN and the mixed features processed by multi-level CNN are input into the improved LSTM model to obtain the error state of the current transformer.

[0053] The present invention also provides at least one computer program product tangibly stored on a non-transitory computer-readable storage medium. The computer program product includes computer executable instructions, such as instructions included in a program module, which are executed in a device on a real or virtual processor of the target to perform the process / method as described above. Typically, program modules include routines, programs, libraries, objects, classes, components, data structures, etc. that perform specific tasks or implement specific abstract data types. In various embodiments, the functions of program modules can be combined or divided between program modules as needed. Machine executable instructions for program modules can be executed in local or distributed devices. In distributed devices, program modules can be located in local and remote storage media.

[0054] The computer program code for implementing the method of the present invention can be written in one or more programming languages. These computer program codes can be provided to the processor of a general-purpose computer, a special-purpose computer or other programmable data processing device, so that the program code, when executed by the computer or other programmable data processing device, causes the function / operation specified in the flow chart and / or block diagram to be implemented. The program code can be executed completely on a computer, partially on a computer, as an independent software package, partially on a computer and partially on a remote computer or completely on a remote computer or server.

[0055] In the context of the present invention, computer program codes or related data may be carried by any appropriate carrier to enable a device, apparatus or processor to perform the various processes and operations described above. Examples of carriers include signals, computer readable media, and the like. Examples of signals may include electrical, optical, radio, acoustic or other forms of propagation signals, such as carrier waves, infrared signals, and the like.

[0056] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with this embodiment can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0057] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A current transformer error state evaluation method based on multi-stage CNN cascade, characterized in that: The following steps are involved: Collect the current time series data of the historical operation of the current transformer, process the time series data, and obtain the time domain image and frequency domain image; A multi-level cascade CNN algorithm is used to extract features from time domain images and frequency domain images respectively to obtain time domain features and frequency domain features, which are then fused to obtain mixed features. The time domain features and frequency domain features processed by the first-level CNN and the mixed features processed by multi-level CNN are input into the improved LSTM model to obtain the error state of the current transformer.

2. The current transformer error state evaluation method based on multi-stage CNN cascade as claimed in claim 1, characterized in that: The current time series data of the historical operation of the current transformer includes zero-sequence unbalanced time series data and negative-sequence unbalanced time series data.

3. The current transformer error state evaluation method based on multi-stage CNN cascade as claimed in claim 1, characterized in that: The time series data is processed to obtain time domain images and frequency domain images, specifically: The Gram angle field is used to convert the time domain signal into a time domain image; Fast Fourier transform is used to convert the time domain signal into frequency domain signal, and Gram angle is used to further convert the frequency domain signal into frequency domain image.

4. The current transformer error state evaluation method based on multi-stage CNN cascade as claimed in claim 1, characterized in that: The time domain features and frequency domain features are fused by cross fusion.

5. The current transformer error state evaluation method based on multi-stage CNN cascade as claimed in claim 4, characterized in that: Up-sampling the time domain features and the frequency domain features to obtain the sampled features; Average pooling and maximum pooling operations are performed on the time domain features and frequency domain features respectively, and the obtained features are subjected to feature splicing operations. Then, after passing through a multi-layer perceptron and an activation function, the time domain feature weights and the frequency domain feature weights are generated. The sampled features are weighted using the generated feature weights to obtain the first time domain weighted features and the first frequency domain weighted features. The first time domain weighted features and the first frequency domain weighted features are fused to obtain mixed features.

6. The current transformer error state evaluation method based on multi-stage CNN cascade as claimed in claim 4, characterized in that: The scale factor of the improved LSTM model is an adaptive scale factor, specifically: ； in, and are the maximum and minimum step sizes determined empirically, , , They are zero-sequence unbalance, negative-sequence unbalance, and the reference value of the sampling current. , , are the parameters that determine the step size adjustment. , , is a hyperparameter, is the variance.

7. A current transformer error state evaluation system based on multi-stage CNN cascade, characterized in that: include: The data acquisition and preprocessing module is configured to: acquire the current time series data of the historical operation of the current transformer, process the time series data, and obtain a time domain image and a frequency domain image; The multi-level cascade module is configured to: use a multi-level cascade CNN algorithm to extract features from the time domain image and the frequency domain image respectively to obtain time domain features and frequency domain features, and fuse the time domain features and the frequency domain features to obtain mixed features; The output module is configured to: input the time domain features and frequency domain features processed by the first-level CNN and the mixed features processed by the multi-level CNN into the improved LSTM model to obtain the error state of the current transformer.

8. An electronic device, characterized in that: The invention comprises a memory and a processor and computer instructions stored in the memory and executed on the processor, wherein the computer instructions are executed by the processor to complete the following steps: Collect the current time series data of the historical operation of the current transformer, process the time series data, and obtain the time domain image and frequency domain image; A multi-level cascade CNN algorithm is used to extract features from time domain images and frequency domain images respectively to obtain time domain features and frequency domain features, which are then fused to obtain mixed features. The time domain features and frequency domain features processed by the first-level CNN and the mixed features processed by multi-level CNN are input into the improved LSTM model to obtain the error state of the current transformer.

9. A computer-readable storage medium, characterized in that: Used to store computer instructions, which, when executed by a processor, implement the following steps: Collect the current time series data of the historical operation of the current transformer, process the time series data, and obtain the time domain image and frequency domain image; A multi-level cascade CNN algorithm is used to extract features from time domain images and frequency domain images respectively to obtain time domain features and frequency domain features, which are then fused to obtain mixed features. The time domain features and frequency domain features processed by the first-level CNN and the mixed features processed by multi-level CNN are input into the improved LSTM model to obtain the error state of the current transformer.

10. A computer program product, characterized in that The invention comprises a computer program, which, when executed by a processor, implements the following steps: Collect the current time series data of the historical operation of the current transformer, process the time series data, and obtain the time domain image and frequency domain image; A multi-level cascade CNN algorithm is used to extract features from time domain images and frequency domain images respectively to obtain time domain features and frequency domain features, which are then fused to obtain mixed features. The time domain features and frequency domain features processed by the first-level CNN and the mixed features processed by multi-level CNN are input into the improved LSTM model to obtain the error state of the current transformer.

Images