Flexible dc transmission and distribution capacitor inner series high square resistance film and preparation method thereof

Abstract

The application discloses a high square resistance film for inner series connection of a flexible direct current transmission and transformation capacitor and a preparation method thereof. The film is prepared by the following steps: treating a polypropylene film to obtain a base film, wherein the treatment comprises stretching, preheating and cooling; forming an aluminum layer and a zinc layer on the base film by a vacuum magnetron sputtering process to obtain a metallized layer; forming a safety film pattern on the metallized layer by a laser cutting process; winding the metallized layer into an element, and spraying a conductive powder on a satin surface of the element to form a gold spraying layer; and assembling the element into a shell, and welding the element with an electric connection wire to obtain the high square resistance film for inner series connection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent preparation, and more particularly, to an inner series high square resistance film for flexible DC power transmission and transformation capacitor and a preparation method thereof. BACKGROUND

[0002] The flexible DC power transmission and transformation capacitor is a power electronic capacitor for VSC-HVDC technology, which can provide energy storage, filtering and voltage stabilization for the DC side of the converter system. The flexible DC power transmission and transformation capacitor usually adopts a metallized film capacitor because it has the advantages of high voltage resistance and high ripple current, low stray inductance and equivalent series resistance, long service life, dry explosion-proof, no polarity, good high-frequency performance, etc.

[0003] The inner series high square resistance film is a new type of metallized safety film, which separates the regular safety film pattern on the insulating material, but the patterns themselves do not cross each other, but are connected to each other through small fuses. When a local weak point of the metal film forms a discharge channel, energy rapidly flows from all directions, and the fuses around the small module of the weak point vaporize, rapidly self-heal without affecting the normal work of the whole metal film. The inner series high square resistance film can improve the withstand voltage and ripple current capacity of the metallized film capacitor, reduce the loss of the fuse part, and work more reliably and safely.

[0004] In the preparation process of the inner series high square resistance film for flexible DC power transmission and transformation capacitor, the key is to ensure that the safety film pattern of the metallized layer meets the predetermined pattern, so that it can play a role in improving the withstand voltage and ripple current capacity of the metallized film capacitor. However, in the existing preparation scheme of the inner series high square resistance film for flexible DC power transmission and transformation capacitor, the performance of the inner series high square resistance film is usually analyzed indirectly to analyze whether the safety film pattern meets the predetermined requirements in the product performance test stage. This detection scheme is lagging.

[0005] Therefore, an optimized preparation method of the inner series high square resistance film for flexible DC power transmission and transformation capacitor is expected. SUMMARY

[0006] In order to solve the above technical problems, the present application is proposed. Embodiments of the present application provide an inner series high square resistance film for flexible DC power transmission and capacitor and a preparation method thereof. The inner series high square resistance film for flexible DC power transmission and capacitor is prepared by treating a polypropylene film to obtain a base film, wherein the treatment includes stretching, preheating and cooling; forming an aluminum layer and a zinc layer on the base film by a vacuum magnetron sputtering process to obtain a metallized layer; forming a safety film pattern on the metallized layer by a laser cutting process; winding the metallized layer into an element, and spraying conductive powder on the satin surface of the element to form a metallized layer; and assembling the element into a shell and welding it with an electrical connection wire to obtain the inner series high square resistance film. In this way, the safety film forming quality is detected during the preparation process to optimize the overall preparation scheme.

[0007] In a first aspect, a preparation method of an inner series high square resistance film for flexible DC power transmission and capacitor is provided, which comprises:

[0008] treating a polypropylene film to obtain a base film, wherein the treatment includes stretching, preheating and cooling;

[0009] forming an aluminum layer and a zinc layer on the base film by a vacuum magnetron sputtering process to obtain a metallized layer;

[0010] forming a safety film pattern on the metallized layer by a laser cutting process;

[0011] winding the metallized layer into an element, and spraying conductive powder on the satin surface of the element to form a metallized layer; and

[0012] assembling the element into a shell and welding it with an electrical connection wire to obtain the inner series high square resistance film.

[0013] In the inner series high square resistance film for flexible DC power transmission and capacitor and the preparation method thereof, the safety film pattern is formed on the metallized layer by a laser cutting process, which includes: obtaining a display image of the safety film pattern collected by a microscope; obtaining a reference design image of the safety film pattern; performing image preprocessing on the display image to obtain a preprocessed display image; inputting the preprocessed display image and the reference design image into a twin detection model comprising a first image encoder and a second image encoder to obtain a display pattern feature vector and a reference pattern feature vector; calculating a transition matrix between the display pattern feature vector and the reference pattern feature vector; inputting the transition matrix into a spatial attention module to obtain a classification feature matrix; and inputting the classification feature matrix into a classifier to obtain a classification result, which is used to represent whether the consistency between the safety film pattern and the design pattern meets a predetermined standard.

[0014] In the inner series high square resistance film for flexible DC power transmission and capacitor and the preparation method thereof, the preprocessing includes image filtering and image enhancement.

[0015] In the above flexible HVDC capacitor inner series high square resistance film and preparation method thereof, the first image encoder and the second image encoder have the same network structure, and the first image encoder and the second image encoder are convolutional neural network models as filters.

[0016] In the above flexible HVDC capacitor inner series high square resistance film and preparation method thereof, the pre-processed display image and the reference design image are input into a twin detection model comprising a first image encoder and a second image encoder to obtain a display pattern feature vector and a reference pattern feature vector, which comprises: performing convolution processing, pooling processing and nonlinear activation processing on the pre-processed display image using the first image encoder of the twin detection model to output the display pattern feature vector from the first image encoder; and performing convolution processing, pooling processing and nonlinear activation processing on the reference design image using the second image encoder of the twin detection model to output the reference pattern feature vector from the second image encoder.

[0017] In the above flexible HVDC capacitor inner series high square resistance film and preparation method thereof, the transition matrix between the display pattern feature vector and the reference pattern feature vector is calculated, which comprises: calculating the Gaussian regression uncertainty factors of the display pattern feature vector and the reference pattern feature vector respectively to obtain a first Gaussian regression uncertainty factor and a second Gaussian regression uncertainty factor; weighting the display pattern feature vector and the reference pattern feature vector respectively with the first Gaussian regression uncertainty factor and the second Gaussian regression uncertainty factor as weights to obtain a weighted display pattern feature vector and a weighted reference pattern feature vector; and calculating the transition matrix of the weighted display pattern feature vector relative to the weighted reference pattern feature vector.

[0018] In the above flexible HVDC capacitor inner series high square resistance film and preparation method thereof, the Gaussian regression uncertainty factors of the display pattern feature vector and the reference pattern feature vector are calculated respectively to obtain a first Gaussian regression uncertainty factor and a second Gaussian regression uncertainty factor, which comprises: calculating the Gaussian regression uncertainty factors of the display pattern feature vector and the reference pattern feature vector with the following optimization formula to obtain a first Gaussian regression uncertainty factor and a second Gaussian regression uncertainty factor; wherein the optimization formula is:

[0019]

[0020]

[0021] wherein V1 represents the display pattern feature vector, V2 represents the reference pattern feature vector, L is the length of the feature vector, μ1 and σ1 2 are the mean and variance of the feature set of each position of the display pattern feature vector, respectively, μ2 and σ2 2 are the mean and variance of the feature set of each position of the reference pattern feature vector, respectively, log is the logarithm with base 2, w1 is the first Gaussian regression uncertainty factor, and w2 is the second Gaussian regression uncertainty factor.

[0022] In the inner series high square resistance film for flexible DC power transmission and preparation method thereof, the transfer matrix is passed through a spatial attention module to obtain a classification feature matrix, including: using a convolutional coding part of the spatial attention module to perform convolutional coding on the transfer matrix to obtain a differential convolution feature matrix; inputting the differential convolution feature matrix into a spatial attention part of the spatial attention module to obtain a spatial attention matrix; passing the spatial attention matrix through a Softmax activation function to obtain a spatial attention feature matrix; and calculating the point multiplication of the spatial attention feature matrix and the differential convolution feature matrix by position to obtain the classification feature matrix.

[0023] In the inner series high square resistance film for flexible DC power transmission and preparation method thereof, the classification feature matrix is passed through a classifier to obtain a classification result, the classification result being used to represent whether the consistency between the safety film pattern and the design pattern meets a predetermined standard, including: expanding the classification feature matrix into a classification feature vector according to a row vector or a column vector; using a plurality of fully connected layers of the classifier to perform fully connected coding on the classification feature vector to obtain an encoded classification feature vector; and passing the encoded classification feature vector through a Softmax classification function of the classifier to obtain the classification result.

[0024] In a second aspect, an inner series high square resistance film for flexible DC power transmission is provided, which is prepared by the preparation method of the inner series high square resistance film for flexible DC power transmission.

[0025] Compared with the prior art, the inner series high square resistance film for flexible DC power transmission and the preparation method thereof provided by the present application process a polypropylene film to obtain a base film, wherein the processing includes stretching, preheating and cooling; an aluminum layer and a zinc layer are formed on the base film by a vacuum magnetron sputtering process to obtain a metallized layer; a safety film pattern is formed on the metallized layer by a laser cutting process; the metallized layer is wound into an element, and a conductive powder is sprayed on the satin surface of the element to form a gold spraying layer; and the element is loaded into a shell and welded with an electrically connected wire to obtain an inner series high square resistance film. In this way, safety film forming quality detection is realized during the preparation process, so as to optimize the overall preparation scheme. Attached Figure Description

[0026] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 This is a schematic diagram illustrating a method for preparing an internally connected high sheet resistance film for a flexible DC transmission and transformation capacitor according to an embodiment of this application.

[0028] Figure 2 This is a flowchart of a method for preparing an internally connected high sheet resistance film for a flexible DC transmission and transformation capacitor according to an embodiment of this application.

[0029] Figure 3 This is a flowchart of a sub-step in step 130 of the method for preparing an internally connected high sheet resistance film for a flexible DC transmission and transformation capacitor according to an embodiment of this application.

[0030] Figure 4 This is a schematic diagram of the structure of step 130 in the method for preparing an internally connected high sheet resistance film for a flexible DC transmission and transformation capacitor according to an embodiment of this application.

[0031] Figure 5 This is a flowchart of a sub-step 134 in the method for preparing an internally connected high sheet resistance film for a flexible DC transmission and transformation capacitor according to an embodiment of this application.

[0032] Figure 6 This is a flowchart of a sub-step 135 in the method for preparing an internally connected high sheet resistance film for a flexible DC transmission and transformation capacitor according to an embodiment of this application.

[0033] Figure 7 This is a flowchart of a sub-step 136 in the method for preparing an internally connected high sheet resistance film for a flexible DC transmission and transformation capacitor according to an embodiment of this application.

[0034] Figure 8 This is a flowchart of a sub-step 137 in the method for preparing an internally connected high sheet resistance film for a flexible DC transmission and transformation capacitor according to an embodiment of this application.

[0035] Figure 9 This is a block diagram of a system for fabricating an internally connected high sheet resistance film for a flexible DC transmission and transformation capacitor according to an embodiment of this application. Detailed Implementation

[0036] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be described. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of the present application.

[0037] Unless otherwise defined, all technical and scientific terms used in the embodiments of the present application have the same meanings as those commonly understood by one of ordinary skill in the art of the present application. The terms used in the present application are only for the purpose of describing the specific embodiments of the present application, and are not intended to limit the scope of the present application.

[0038] In the embodiments of the present application, it should be noted that unless otherwise specified and limited, the term "connection" should be understood broadly, for example, it can be an electrical connection, or a communication between two elements, it can be directly connected, or indirectly connected through an intermediate medium. Those of ordinary skill in the art can understand the specific meaning of the above-mentioned term according to the specific circumstances.

[0039] It should be noted that the terms "first", "second", "third" in the embodiments of the present application are only to distinguish similar objects, and do not represent a specific order of the objects. Understandably, "first", "second", "third" can be interchanged in specific order or sequence as allowed. It should be understood that the objects distinguished by "first", "second", "third" can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein.

[0040] Specifically, the present application first provides a preparation method of an inner series high square resistance film for flexible HVDC capacitor, which comprises the following steps: S1, treating a polypropylene film to obtain a base film, wherein the treatment comprises stretching, preheating and cooling; S2, forming an aluminum layer and a zinc layer on the base film by a vacuum magnetron sputtering process to obtain a metallized layer; S3, forming a safety film pattern on the metallized layer by a laser cutting process; S4, winding the metallized layer into an element, and spraying conductive powder on the satin surface of the element to form a gold spraying layer; and S5, assembling the element into a shell, and welding with an electrical connection wire to obtain an inner series high square resistance film.

[0041] For the technical problems proposed in the background art, the technical concept of the present application is to collect the display image of the safety film pattern by a microscope, and compare and analyze the display image with the reference design image of the safety film pattern to determine whether the consistency between the safety film pattern and the design pattern meets the predetermined standard. In this way, safety film forming quality detection is realized during the preparation process to optimize the overall preparation scheme.

[0042] Specifically, in the technical solution of the present application, first, a display image of the security film pattern collected by a microscope is acquired, and a reference design image of the security film pattern is acquired. It should be understood that during the manufacturing process of the security film, microscopic observation of the security film is usually required to ensure its quality and consistency. The microscope can magnify and capture the details and lines of the surface of the security film. The reference design image of the security film pattern is acquired because when judging the consistency between the security film pattern and the design pattern, it is necessary to compare the security film pattern with its corresponding reference design image to determine whether they meet the predetermined standard. The reference design image is usually a theoretical template or standard pattern provided by a designer according to the manufacturing requirements and specifications of the security film. Therefore, the reference design image can be used as a "standard" or "golden sample" for the quality of the security film pattern and is used for comparison with the actually collected security film pattern to detect possible errors and deviations.

[0043] Next, the display image is pre-processed to obtain a pre-processed display image. Here, the display image is pre-processed to obtain a pre-processed display image because when judging the consistency between the security film pattern and the design pattern, unnecessary factors such as noise and interference need to be reduced. In the technical solution of the present application, the pre-processing usually includes image filtering and image enhancement, wherein the image filtering can remove high-frequency noise and low-frequency noise in the image, making the image smoother and clearer. For example, methods such as Gaussian filtering, median filtering, etc. can be used to eliminate noise and artifacts, etc.; and the image enhancement can enhance the contrast and brightness of the image, etc., making the image more recognizable and distinguishable, for example, methods such as histogram equalization, gray stretch, etc. can be used to adjust the color saturation and brightness of the image, etc.

[0044] Further, the pre-processed display image and the reference design image are input into a twin detection model comprising a first image encoder and a second image encoder to obtain a display pattern feature vector and a reference pattern feature vector, wherein the first image encoder and the second image encoder have the same network structure. That is, in the technical solution of the present application, the twin detection model comprising the first image encoder and the second image encoder is used to extract image features from the pre-processed display image and the reference design image to map the pre-processed display image and the reference design image into a high-dimensional feature space to obtain the display pattern feature vector and the reference pattern feature vector.

[0045] In particular, in the technical solution of the present application, the first image encoder and the second image encoder included in the twin detection model have the same network structure, so that the inconsistent parts of the pre-processed display image and the reference design image can be captured and amplified, and then by comparing the display pattern feature vector and the reference pattern feature vector, whether the consistency between the security film pattern and the design pattern meets the predetermined standard can be analyzed.

[0046] Specifically, in the technical solution of the present application, the transition matrix between the display pattern feature vector and the reference pattern feature vector is further calculated. The reason for calculating the transition matrix between the display pattern feature vector and the reference pattern feature vector is that when judging the consistency between the security film pattern and the design pattern, the similarity between them needs to be quantified as a numerical value, and whether they meet the predetermined standard is determined through these numerical values. The transition matrix can establish a relationship between the two feature vectors and map them into a common space, so as to better compare the similarity between them.

[0047] In particular, in the technical solution of the present application, considering that the eigenvalues of each position in the transition matrix are used to represent the position-by-position correlation information of the display pattern feature vector and the reference pattern feature vector, and the eigenvalues of each element in the transition matrix have different contributions to the final classification judgment. In order to fully utilize the weight heterogeneity of spatial dimensions, in the technical solution of the present application, the transition matrix is passed through a spatial attention module to obtain a classification feature matrix.

[0048] Further, the classification feature matrix is passed through a classifier to obtain a classification result, which is used to represent whether the consistency between the security film pattern and the design pattern meets the predetermined standard. The classifier is a model based on machine learning, which can classify the input feature vector. The classifier is usually obtained from a training set, and by learning the relationship between different feature vectors and categories in the training set, the classification of new feature vectors is realized. By using the classifier, the similarity between the security film pattern and the design pattern can be well distinguished, so as to judge whether they meet the predetermined standard. The classification result can be binary (consistent / inconsistent) or multi-element (A / B / C / D, etc.), and different evaluation indicators can be used to measure the performance and accuracy of the classifier. By comparing the classification result and the standard value, whether the consistency between the security film pattern and the design pattern meets the predetermined standard can be concluded.

[0049] In particular, when calculating the transition matrix between the display pattern feature vector and the reference pattern feature vector, considering that the source image noise of the display image and the reference design image cannot be guaranteed to be the same image semantic encoding direction in the image semantic feature encoding process by the first image encoder and the second image encoder, which will cause the existence of regression error in the transition matrix as the image semantic feature domain transition expression between the display pattern feature vector and the reference pattern feature vector, and the regression error in the classification feature matrix obtained by the transition matrix through the spatial attention module, affecting the accuracy of the classification result obtained by the classification feature matrix through the classifier.

[0050] Based on this, in the technical solution of the present application, the Gaussian regression uncertainty factors of the display pattern feature vector V1 and the reference pattern feature vector V2 are calculated respectively, denoted as:

[0051]

[0052]

[0053] L is the length of the feature vector, μ1 and σ1 2 are the mean and variance of the feature set v 1i ∈V1, and μ2 and σ2 2 are the mean and variance of the feature set v 2i ∈V2, and log is the logarithm with base 2.

[0054] Here, for the agnostic regression of the classification feature matrix caused by the distribution uncertainty information of the integrated feature set of the display pattern feature vector V1 and the reference pattern feature vector V2, the mean and variance as statistical quantization parameters are used to perform scalar measurement of the statistical characteristics of the feature set, so as to extend the normal distribution recognition mode of the feature representation of the source image noise to the unknown distribution regression mode, realize the transfer learning based on natural distribution transition in the feature set scale, and then calculate the transition matrix of the display pattern feature vector and the reference pattern feature vector by weighting them with the above Gaussian regression uncertainty factors. Thus, the self-calibration uncertainty correction of the display pattern feature vector and the reference pattern feature vector can be realized, so as to correct the regression error existing in the transition matrix and the classification feature matrix obtained from the transition matrix, and improve the accuracy of the classification result obtained by the classification feature matrix through the classifier.

[0055] Figure 1This is a schematic diagram illustrating a method for fabricating an internally connected high sheet resistance film for a flexible DC transmission and transformation capacitor according to an embodiment of this application. Figure 1 As shown, in this application scenario, firstly, the image is obtained from a microscope (e.g., as shown in the image). Figure 1 The image of the safety film pattern captured by M (as shown) (e.g., as shown) Figure 1 The C1 shown in the diagram), and a reference design image of the safety film pattern (e.g., as shown in the diagram). Figure 1 (as shown in C2); then, the acquired display image of the safety film pattern and the reference design image of the safety film pattern are input to a server (e.g., such as...) that has deployed an algorithm for fabricating a flexible DC transmission and transformation capacitor with an internally connected high sheet resistance film. Figure 1 In the S shown, the server is capable of processing the display image of the safety film pattern and the reference design image of the safety film pattern based on the fabrication algorithm of the flexible DC transmission and transformation capacitor with an internally connected high sheet resistance film, to generate a classification result indicating whether the consistency between the safety film pattern and the design pattern meets a predetermined standard.

[0056] After introducing the basic principles of this application, various non-limiting embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0057] In one embodiment of this application, Figure 2 This is a flowchart illustrating a method for preparing an internally connected high sheet resistance film for a flexible DC transmission and transformation capacitor according to an embodiment of this application. Figure 2 As shown, a method 100 for preparing an internally connected high sheet resistance film for a flexible DC transmission and transformation capacitor according to an embodiment of this application includes: 110, processing a polypropylene film to obtain a base film, wherein the processing includes stretching, preheating, and cooling; 120, forming an aluminum layer and a zinc layer on the base film by a vacuum magnetron sputtering process to obtain a metallization layer; 130, forming a safety film pattern on the metallization layer by a laser cutting process; 140, winding the metallization layer into an element and spraying conductive powder onto the satin surface of the element to form a gold-plated layer; and 150, assembling the element into a housing and welding it to an electrical connection wire to obtain an internally connected high sheet resistance film.

[0058] Specifically, the application first provides a preparation method of an inner series high square resistance film for a flexible HVDC capacitor, comprising the following steps: S1, treating a polypropylene film to obtain a base film, wherein the treatment comprises stretching, preheating and cooling; S2, forming an aluminum layer and a zinc layer on the base film by a vacuum magnetron sputtering process to obtain a metallized layer; S3, forming a security film pattern on the metallized layer by a laser cutting process; S4, winding the metallized layer into an element, and spraying conductive powder on the satin surface of the element to form a gold spraying layer; and S5, assembling the element into a shell, and welding connection with an electrically connected wire to obtain the inner series high square resistance film.

[0059] Figure 3 The flow chart of the sub-steps of step 130 in the preparation method of the inner series high square resistance film for the flexible HVDC capacitor according to the embodiment of the application is shown in FIG. 13. Figure 3 As shown in FIG. 13, the security film pattern is formed on the metallized layer by a laser cutting process, comprising: 131, acquiring a display image of the security film pattern collected by a microscope; 132, acquiring a reference design image of the security film pattern; 133, performing image preprocessing on the display image to obtain a preprocessed display image; 134, passing the preprocessed display image and the reference design image through a twin detection model comprising a first image encoder and a second image encoder to obtain a display pattern feature vector and a reference pattern feature vector; 135, calculating a transition matrix between the display pattern feature vector and the reference pattern feature vector; 136, passing the transition matrix through a spatial attention module to obtain a classification feature matrix; and 137, passing the classification feature matrix through a classifier to obtain a classification result, wherein the classification result is used to represent whether the consistency between the security film pattern and the design pattern meets a predetermined standard.

[0060] Figure 4 The architectural schematic diagram of step 130 in the preparation method of the inner series high square resistance film for the flexible HVDC capacitor according to the embodiment of the application is shown in FIG. 14. Figure 3As shown, in the network architecture, first, a display image of the security film pattern collected by a microscope is acquired; then, a reference design image of the security film pattern is acquired; next, the display image is pre-processed to obtain a pre-processed display image; then, the pre-processed display image and the reference design image are passed through a twin detection model comprising a first image encoder and a second image encoder to obtain a display pattern feature vector and a reference pattern feature vector; next, a transition matrix between the display pattern feature vector and the reference pattern feature vector is calculated; then, the transition matrix is passed through a spatial attention module to obtain a classification feature matrix; and finally, the classification feature matrix is passed through a classifier to obtain a classification result, which is used to represent whether the consistency between the security film pattern and the design pattern meets a predetermined standard.

[0061] Specifically, in steps 131 and 132, a display image of the security film pattern collected by a microscope is acquired; and a reference design image of the security film pattern is acquired. For the technical problems proposed in the background art, the technical concept of the present application is to collect the display image of the security film pattern by a microscope, and compare and analyze the display image with the reference design image of the security film pattern to determine whether the consistency between the security film pattern and the design pattern meets a predetermined standard, so as to realize security film forming quality detection during preparation to optimize the overall preparation scheme.

[0062] Specifically, in the technical solution of the present application, first, a display image of the security film pattern collected by a microscope is acquired, and a reference design image of the security film pattern is acquired. It should be understood that during the manufacturing process of the security film, it is usually necessary to observe the security film under a microscope to ensure its quality and consistency. The microscope can magnify and capture the details and lines on the surface of the security film. The reference design image of the security film pattern is acquired because when determining the consistency between the security film pattern and the design pattern, it is necessary to compare the security film pattern with its corresponding reference design image to determine whether they meet the predetermined standard. The reference design image is usually a theoretical template or standard pattern provided by a designer according to the manufacturing requirements and specifications of the security film, therefore, the reference design image can be used as a "standard" or "golden sample" for the quality of the security film pattern, and is used to compare with the actually collected security film pattern, so as to detect possible errors and deviations.

[0063] Specifically, in step 133, the display image is image pre-processed to obtain a pre-processed display image. Then, the display image is image pre-processed to obtain a pre-processed display image. Here, the display image is image pre-processed to obtain a pre-processed display image because when judging the consistency between the security film pattern and the design pattern, it is necessary to reduce unnecessary factors such as noise and interference.

[0064] In the technical solution of the present application, the pre-processing generally includes image filtering and image enhancement, wherein the image filtering can remove high-frequency noise and low-frequency noise in the image, making the image smoother and clearer. For example, methods such as Gaussian filtering and median filtering can be used to eliminate noise and artifacts; and the image enhancement can enhance the contrast and brightness of the image and other features, making the image more recognizable and distinguishable, for example, methods such as histogram equalization and gray scale stretching can be used to adjust the color saturation and brightness of the image and other parameters.

[0065] Specifically, in step 134, the pre-processed display image and the reference design image are passed through a Siamese detection model comprising a first image encoder and a second image encoder to obtain a display pattern feature vector and a reference pattern feature vector. Further, the pre-processed display image and the reference design image are passed through a Siamese detection model comprising a first image encoder and a second image encoder to obtain a display pattern feature vector and a reference pattern feature vector, wherein the first image encoder and the second image encoder have the same network structure. That is, in the technical solution of the present application, the pre-processed display image and the reference design image are image feature extracted using the Siamese detection model comprising a first image encoder and a second image encoder to map the pre-processed display image and the reference design image into a high-dimensional feature space to obtain the display pattern feature vector and the reference pattern feature vector.

[0066] In particular, in the technical solution of the present application, the first image encoder and the second image encoder included in the Siamese detection model have the same network structure, so that the inconsistent parts of the pre-processed display image and the reference design image are captured and amplified, and then by comparing the display pattern feature vector and the reference pattern feature vector, it can be analyzed whether the consistency between the security film pattern and the design pattern meets the predetermined standard.

[0067] Wherein, the first image encoder and the second image encoder have the same network structure, and the first image encoder and the second image encoder are convolutional neural network models as filters.

[0068] Figure 5For the flow chart of the sub-step of step 134 in the preparation method of the inner series high square resistance film for flexible HVDC power transmission and transformation capacitor according to the embodiment of the application, as shown in Figure 5 As shown in the flow chart of the sub-step of step 134 in the preparation method of the inner series high square resistance film for flexible HVDC power transmission and transformation capacitor according to the embodiment of the application, as shown in

[0069] A convolutional neural network (CNN) is an artificial neural network and has a wide range of applications in the field of image recognition. The convolutional neural network can include an input layer, a hidden layer and an output layer, wherein the hidden layer can include a convolution layer, a pooling layer, an activation layer and a full connection layer, etc. The previous layer performs corresponding operations according to the input data and outputs the operation result to the next layer. The initial data is obtained after the operation of multiple layers.

[0070] The convolutional neural network model uses a convolution kernel as a feature filtering factor and has very excellent performance in image local feature extraction. Compared with traditional image feature extraction algorithms based on statistics or feature engineering, the convolutional neural network model has stronger feature extraction generalization ability and fitting ability.

[0071] Specifically, in step 135, the transition matrix between the display pattern feature vector and the reference pattern feature vector is calculated. Specifically, in the technical solution of the present application, the transition matrix between the display pattern feature vector and the reference pattern feature vector is further calculated. The transition matrix between the display pattern feature vector and the reference pattern feature vector is calculated because when judging the consistency between the security film pattern and the design pattern, it is necessary to quantify the similarity between them into a numerical value, and determine whether they meet the predetermined standard through these numerical values. The transition matrix can establish a relationship between the two feature vectors and map them into a common space in order to better compare the similarity between them.

[0072] In particular, in the technical solution of the present application, the eigenvalues of each position in the transition matrix are used to represent the position-by-position correlation information of the display pattern feature vector and the reference pattern feature vector, and the eigenvalues of each element in the transition matrix have different contribution degrees to the final classification judgment.

[0073] Figure 6 The flowchart of the sub-step of step 135 in the preparation method of the inner series high square resistance film for flexible HVDC capacitor according to the embodiment of the present application is shown in Figure 6 As shown in the figure, the transition matrix between the display pattern feature vector and the reference pattern feature vector is calculated, including: 1351, the Gaussian regression uncertainty factors of the display pattern feature vector and the reference pattern feature vector are calculated respectively to obtain first and second Gaussian regression uncertainty factors; 1352, the first and second Gaussian regression uncertainty factors are used as weights to weight the display pattern feature vector and the reference pattern feature vector respectively to obtain a weighted display pattern feature vector and a weighted reference pattern feature vector; and 1353, the transition matrix of the weighted display pattern feature vector relative to the weighted reference pattern feature vector is calculated.

[0074] In particular, when calculating the transition matrix between the display pattern feature vector and the reference pattern feature vector, it is considered that the source image noise of the display image and the reference design image cannot be guaranteed to have the same image semantic encoding direction in the image semantic feature encoding process by the first image encoder and the second image encoder, which will lead to the existence of regression errors in the transition matrix as the image semantic feature domain transition expression between the display pattern feature vector and the reference pattern feature vector, which are superimposed by the Gaussian distribution uncertainty of the display pattern feature vector and the reference pattern feature vector respectively, so that the classification feature matrix obtained by the spatial attention module from the transition matrix also has regression errors, which affects the accuracy of the classification result obtained by the classifier from the classification feature matrix.

[0075] Based on this, in the technical solution of the present application, the Gaussian regression uncertainty factors of the display pattern feature vector V1 and the reference pattern feature vector V2 are calculated respectively, which are represented as: the Gaussian regression uncertainty factors of the display pattern feature vector and the reference pattern feature vector are calculated by the following optimization formula to obtain first and second Gaussian regression uncertainty factors; wherein the optimization formula is:

[0076]

[0077]

[0078] wherein, V1 represents the display pattern feature vector, V2 represents the reference pattern feature vector, L is the length of the feature vector, μ1 and σ1 2 are the mean and variance of the feature set of each position of the display pattern feature vector respectively, μ2 and σ2 2 are the mean and variance of the feature set of each position of the reference pattern feature vector respectively, log is the logarithm with base 2, w1 is the first Gaussian regression uncertainty factor, w2 is the second Gaussian regression uncertainty factor.

[0079] Here, in order to avoid the agnostic regression of the classification feature matrix caused by the distribution uncertainty information of the integrated feature set of the display pattern feature vector V1 and the reference pattern feature vector V2 respectively, the scalar measurement of the statistical characteristics of the feature set is performed by using the mean and variance as the statistical quantization parameters, so as to expand the normal distribution cognition mode of the feature representation of the source image noise to the unknown distribution regression mode, and realize the transfer learning based on the natural distribution transfer in the feature set scale. Thus, by weighting the display pattern feature vector and the reference pattern feature vector respectively by using the above Gaussian regression uncertainty factors and then calculating the transfer matrix, the self-calibration uncertainty correction of the display pattern feature vector and the reference pattern feature vector can be realized, so as to correct the regression error existing in the transfer matrix and the classification feature matrix obtained from the transfer matrix, and improve the accuracy of the classification result obtained by the classifier through the classification feature matrix.

[0080] Specifically, in step 136, the transfer matrix is passed through a spatial attention module to obtain a classification feature matrix. In order to make full use of the weight heterogeneity in the spatial dimension, in the technical solution of the present application, the transfer matrix is passed through a spatial attention module to obtain a classification feature matrix.

[0081] Figure 7 The flow chart of the sub-step of step 136 in the preparation method of the inner series high sheet resistance film for flexible HVDC power transmission and transformation capacitor according to the embodiment of the present application is shown as follows, Figure 7 The transfer matrix is passed through a spatial attention module to obtain a classification feature matrix, including: 1361, using the convolutional encoding part of the spatial attention module to convolutionally encode the transfer matrix to obtain a differential convolution feature matrix; 1362, inputting the differential convolution feature matrix into the spatial attention part of the spatial attention module to obtain a spatial attention matrix; 1363, passing the spatial attention matrix through a Softmax activation function to obtain a spatial attention feature matrix; and 1364, calculating the point-by-position multiplication of the spatial attention feature matrix and the differential convolution feature matrix to obtain the classification feature matrix.

[0082] Attention mechanism is a data processing method in machine learning, widely used in natural language processing, image recognition and speech recognition, etc. On the one hand, attention mechanism is to enable the network to automatically learn the places that need to be paid attention to in the picture or text sequence; on the other hand, attention mechanism generates a mask through the operation of neural network, and the weight of the value on the mask. Generally speaking, the spatial attention mechanism takes the average of different channels of the same pixel point, and then obtains the spatial feature through some convolution and upsampling operations. The pixel points of each layer of the spatial feature are assigned different weights.

[0083] Specifically, in step 137, the classification feature matrix is passed through a classifier to obtain a classification result, which indicates whether the consistency between the security film pattern and the design pattern meets the predetermined standard. Further, the classification feature matrix is passed through a classifier to obtain a classification result, which indicates whether the consistency between the security film pattern and the design pattern meets the predetermined standard. The classifier is a model based on machine learning, which can classify the input feature vector.

[0084] The classifier is usually obtained from a training set and learns the relationship between different feature vectors and categories in the training set to classify new feature vectors. By using the classifier, the similarity between the security film pattern and the design pattern can be well distinguished to determine whether they meet the predetermined standard. The classification result can be binary (consistent / inconsistent) or multi-valued (A / B / C / D, etc.), and different evaluation indicators can be used to measure the performance and accuracy of the classifier. By comparing the classification result and the standard value, it can be concluded whether the consistency between the security film pattern and the design pattern meets the predetermined standard.

[0085] Figure 8 For the flowchart of the sub-step of step 137 in the preparation method of the inner series high sheet resistance film for flexible HVDC power transmission and transformation capacitor according to the embodiment of the present application, as shown in Figure 8 The classification feature matrix is passed through a classifier to obtain a classification result, which indicates whether the consistency between the security film pattern and the design pattern meets the predetermined standard, including: 1371, the classification feature matrix is expanded into a classification feature vector according to the row vector or the column vector; 1372, the classification feature vector is fully connected and coded using multiple fully connected layers of the classifier to obtain an encoded classification feature vector; and 1373, the encoded classification feature vector is passed through the Softmax classification function of the classifier to obtain the classification result.

[0086] In summary, the preparation method 100 of the inner series high square resistance film for the flexible DC power transmission and transformation capacitor according to the embodiments of the present application is illustrated, which processes a polypropylene film to obtain a base film, wherein the processing includes stretching, preheating and cooling; forms an aluminum layer and a zinc layer on the base film by a vacuum magnetron sputtering process to obtain a metallized layer; forms a safety film pattern on the metallized layer by a laser cutting process; winds the metallized layer into an element, and sprays conductive powder on the satin surface of the element to form a metallized layer; and packs the element into a shell, and is welded with an electrical connection wire to obtain an inner series high square resistance film. In this way, the safety film forming quality detection is realized in the preparation process, so as to optimize the overall preparation scheme.

[0087] In an embodiment of the present application, an inner series high square resistance film for the flexible DC power transmission and transformation capacitor is also provided, which is prepared by the preparation method of the inner series high square resistance film for the flexible DC power transmission and transformation capacitor.

[0088] In an embodiment of the present application, Figure 9 A block diagram of a preparation system of the inner series high square resistance film for the flexible DC power transmission and transformation capacitor according to the embodiments of the present application is shown. As shown in the figure, Figure 9 The preparation system 200 of the inner series high square resistance film for the flexible DC power transmission and transformation capacitor according to the embodiments of the present application includes: a film processing module 210 for processing a polypropylene film to obtain a base film, wherein the processing includes stretching, preheating and cooling; a vacuum magnetron sputtering module 220 for forming an aluminum layer and a zinc layer on the base film by a vacuum magnetron sputtering process to obtain a metallized layer; a laser cutting module 230 for forming a safety film pattern on the metallized layer by a laser cutting process; a satin surface spraying module 240 for winding the metallized layer into an element, and spraying conductive powder on the satin surface of the element to form a metallized layer; and an electrical connection module 250 for packing the element into a shell, and welding with an electrical connection wire to obtain an inner series high square resistance film.

[0089] In one specific example, in the above preparation system of the high square resistance film for internal series connection of flexible HVDC power transmission and transformation capacitor, the laser cutting module comprises: a display image acquisition unit 231 configured to acquire a display image of the security film pattern collected by a microscope; a reference image acquisition unit 232 configured to acquire a reference design image of the security film pattern; an image preprocessing unit 233 configured to perform image preprocessing on the display image to obtain a preprocessed display image; a twin detection unit 234 configured to pass the preprocessed display image and the reference design image through a twin detection model comprising a first image encoder and a second image encoder to obtain a display pattern feature vector and a reference pattern feature vector; a transfer matrix calculation unit 235 configured to calculate a transfer matrix between the display pattern feature vector and the reference pattern feature vector; a spatial attention unit 236 configured to pass the transfer matrix through a spatial attention module to obtain a classification feature matrix; and a classification result generation unit 237 configured to pass the classification feature matrix through a classifier to obtain a classification result, wherein the classification result is used to indicate whether the consistency between the security film pattern and the design pattern meets a predetermined standard.

[0090] In one specific example, in the above preparation system of the high square resistance film for internal series connection of flexible HVDC power transmission and transformation capacitor, the preprocessing comprises image filtering and image enhancement.

[0091] In one specific example, in the above preparation system of the high square resistance film for internal series connection of flexible HVDC power transmission and transformation capacitor, the first image encoder and the second image encoder have the same network structure, and the first image encoder and the second image encoder are convolutional neural network models as filters.

[0092] In one specific example, in the above preparation system of the high square resistance film for internal series connection of flexible HVDC power transmission and transformation capacitor, the twin detection unit comprises: a first image encoding subunit configured to perform convolution processing, pooling processing and nonlinear activation processing on the preprocessed display image using a first image encoder of the twin detection model to output the display pattern feature vector from the first image encoder; and a second image encoding subunit configured to perform convolution processing, pooling processing and nonlinear activation processing on the reference design image using a second image encoder of the twin detection model to output the reference pattern feature vector from the second image encoder.

[0093] In one specific example, in the preparation system for the high square resistance film for inner series connection of the flexible DC power transmission and transformation capacitor, the transfer matrix calculation unit comprises: an uncertainty factor calculation subunit, configured to calculate Gaussian regression uncertainty factors of the display pattern feature vector and the reference pattern feature vector respectively to obtain a first Gaussian regression uncertainty factor and a second Gaussian regression uncertainty factor; a weighting subunit, configured to weight the display pattern feature vector and the reference pattern feature vector respectively by taking the first Gaussian regression uncertainty factor and the second Gaussian regression uncertainty factor as weights to obtain a weighted display pattern feature vector and a weighted reference pattern feature vector; and a calculation subunit, configured to calculate a transfer matrix of the weighted display pattern feature vector relative to the weighted reference pattern feature vector.

[0094] In one specific example, in the preparation system for the high square resistance film for inner series connection of the flexible DC power transmission and transformation capacitor, the uncertainty factor calculation subunit is configured to calculate Gaussian regression uncertainty factors of the display pattern feature vector and the reference pattern feature vector by using an optimization formula to obtain a first Gaussian regression uncertainty factor and a second Gaussian regression uncertainty factor, wherein the optimization formula is as follows:

[0095]

[0096]

[0097] wherein V1 represents the display pattern feature vector, V2 represents the reference pattern feature vector, L is the length of the feature vector, μ1 and σ1 2 are the mean and variance of the feature set at each position of the display pattern feature vector, μ2 and σ2 2 are the mean and variance of the feature set at each position of the reference pattern feature vector, log is the logarithm with base 2, w1 is the first Gaussian regression uncertainty factor, and w2 is the second Gaussian regression uncertainty factor.

[0098] In one specific example, in the preparation system for the high square resistance film for inner series connection of the flexible DC power transmission and transformation capacitor, the spatial attention unit comprises: a convolution subunit, configured to perform convolutional encoding on the transfer matrix by using a convolutional encoding part of the spatial attention module to obtain a differential convolution feature matrix; an attention subunit, configured to input the differential convolution feature matrix into a spatial attention part of the spatial attention module to obtain a spatial attention matrix; an activation subunit, configured to pass the spatial attention matrix through a Softmax activation function to obtain a spatial attention feature matrix; and a point multiplication calculation subunit, configured to calculate the positional point multiplication of the spatial attention feature matrix and the differential convolution feature matrix to obtain the classification feature matrix.

[0099] In one specific example, in the preparation system of the high square resistance film for inner series connection of flexible DC power transmission and transformation capacitor, the classification result generation unit comprises: an arrangement sub-unit, configured to expand the classification feature matrix into a classification feature vector according to a row vector or a column vector; a fully connected coding sub-unit, configured to perform fully connected coding on the classification feature vector using a plurality of fully connected layers of the classifier to obtain a coded classification feature vector; and a classification sub-unit, configured to pass the coded classification feature vector through a Softmax classification function of the classifier to obtain the classification result.

[0100] Here, those skilled in the art can understand that the specific functions and operations of each unit and module in the preparation system of the high square resistance film for inner series connection of flexible DC power transmission and transformation capacitor have been described in detail above with reference to the preparation method of the high square resistance film for inner series connection of flexible DC power transmission and transformation capacitor Figures 1 to 8 , and therefore, the repeated description thereof will be omitted.

[0101] The present application also provides a computer program product, which comprises instructions that, when executed, cause an apparatus to perform operations corresponding to the above-described method.

[0102] In an embodiment of the present application, a computer-readable storage medium storing a computer program for executing the above-described method is also provided.

[0103] It should be understood that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, it can be provided in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, it can be provided in the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0104] The method, system, and computer program product of the embodiments of the present application are described in the form of flowcharts and / or block diagrams. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts 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, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one flow or a plurality of flows and / or blocks Figure 1 The functions specified in one flow or a plurality of flows and / or blocks

[0105] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the flow Figure 1 one or more flowcharts and / or blocks Figure 1 one or more flowcharts and / or blocks

[0106] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the flow Figure 1 one or more flowcharts and / or blocks Figure 1 one or more flowcharts and / or blocks

[0107] The above description sets forth numerous specific details to provide a thorough understanding of the application. However, those of skill in the art will appreciate that the application can be practiced without these specific details. In other instances, well-known methods, procedures, components, and circuits have not been described in detail since not to unnecessarily obscure aspects of the application. The above description is intended to be illustrative, and not restrictive. Many other embodiments will be apparent to those of skill in the art upon reviewing the above description. The scope of the application should, therefore, be determined not with reference to the above description, but instead with reference to the appended claims, along with their full scope of equivalents.

[0108] The block diagrams of the devices, apparatus, systems, and flowcharts herein are meant to be exemplary only and are not intended to suggest any requirement on the connection, arrangement, and configuration in the practice thereof. As will be realized by one of ordinary skill in the art, the devices, apparatus, systems, and flowcharts can be connected, arranged, and configured in a wide variety of ways. The phraseology and terminology used herein is for the purpose of description and should not be regarded as limiting. The use of "including," "comprising," "having," "containing," or "housing" and variations thereof herein is intended to be broad and encompass the option of zero, one, or more steps, components, or units. The use of "or" herein is meant to be the inclusive, and not the exclusive use. The use of "such as," "for example," and "for instance" herein is meant to be the inclusive, and not the exclusive use. Only claims may limit the scope of this application.

[0109] It is also noted that the specific order or hierarchy of steps, operations, or processes in the processes disclosed herein are an illustration only and can be performed in many other orders or hierarchies. The specific steps, operations, or processes represented herein are merely illustrative, and should not be considered a restriction. Any and all combinations of claims may lead to one or more patent applications.

[0110] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein can be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0111] Finally, it is also noted that the illustrative language used herein is merely intended to be illustrative of the present application and is not intended to limit the scope of the application as claimed. Using "for example" or "as an example" merely indicates one of the possible embodiments of the application. Thus, the scope of the present application is not intended to be limited to the illustrative examples given herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0112] The above description has been presented for the purpose of illustration and description. Furthermore, this description is not intended to limit the embodiments of the application to forms disclosed herein. Although various example aspects and embodiments have been discussed above, those of ordinary skill in the art will appreciate a variety of modifications, alternatives, permutations, additions, and sub-combinations, which fall within the scope of the application.

Claims

1. A method for preparing an inner series high square resistance film for a flexible HVDC power transmission and transformation capacitor, characterized by, The method comprises: treating a polypropylene film to obtain a base film, wherein the treatment comprises stretching, preheating and cooling; forming an aluminum layer and a zinc layer on the base film by a vacuum magnetron sputtering process to obtain a metallized layer; forming a security film pattern on the metallized layer by a laser cutting process; winding the metallized layer into an element, and spraying a conductive powder on the satin surface of the element to form a metallized layer; and assembling the element into a housing and welding it with an electrical connection wire to obtain an inner string high square resistance film; forming a security film pattern on the metallized layer by a laser cutting process, comprising: obtaining a display image of the security film pattern collected by a microscope; obtaining a reference design image of the security film pattern; image preprocessing the display image to obtain a preprocessed display image; passing the preprocessed display image and the reference design image through a twin detection model comprising a first image encoder and a second image encoder to obtain a display pattern feature vector and a reference pattern feature vector; calculating the transition matrix between the display pattern feature vector and the reference pattern feature vector; passing the transition matrix through a spatial attention module to obtain a classification feature matrix; and passing the classification feature matrix through a classifier to obtain a classification result, which is used to represent whether the consistency between the security film pattern and the design pattern meets a predetermined standard; calculating the transition matrix between the display pattern feature vector and the reference pattern feature vector, comprising: respectively calculating the Gaussian regression uncertainty factors of the display pattern feature vector and the reference pattern feature vector to obtain a first Gaussian regression uncertainty factor and a second Gaussian regression uncertainty factor; respectively weighting the display pattern feature vector and the reference pattern feature vector with the first Gaussian regression uncertainty factor and the second Gaussian regression uncertainty factor as weights to obtain a weighted display pattern feature vector and a weighted reference pattern feature vector; and calculating the transition matrix of the weighted display pattern feature vector relative to the weighted reference pattern feature vector; calculating the Gaussian regression uncertainty factors of the display pattern feature vector and the reference pattern feature vector to obtain a first Gaussian regression uncertainty factor and a second Gaussian regression uncertainty factor with the following optimization formula: wherein the optimization formula is: wherein, represents the display pattern feature vector, represents the reference pattern feature vector, is the length of the feature vector, and are the mean and variance of the feature set of each position of the display pattern feature vector, respectively, and are the mean and variance of the feature set of each position of the reference pattern feature vector, respectively, is the base 2 logarithm, is a first Gaussian regression uncertainty factor, is a second Gaussian regression uncertainty factor.

2. The method of claim 1, wherein the method is used for the preparation of an inner series high-squarce-resistance film for a flexible HVDC capacitor. The preprocessing includes image filtering and image enhancement.

3. The method of claim 2, wherein the method is performed in a vacuum chamber. The first image encoder and the second image encoder have the same network structure, and the first image encoder and the second image encoder are convolutional neural network models as filters.

4. The method of claim 3, wherein the method is performed in a vacuum chamber. passing the preprocessed display image and the reference design image through a twin detection model comprising a first image encoder and a second image encoder to obtain a display pattern feature vector and a reference pattern feature vector, comprising: convolution processing, pooling processing and nonlinear activation processing of the preprocessed display image using the first image encoder of the twin detection model to output the display pattern feature vector from the first image encoder; and The second image encoder using the twin detection model is used to perform convolution processing, pooling processing and nonlinear activation processing on the reference design image to output the reference pattern feature vector.

5. The method of claim 4, wherein the method is performed in a vacuum chamber. The transfer matrix is input into a spatial attention module to obtain a classification feature matrix, including: The transfer matrix is convolutionally encoded using a convolutional encoding part of the spatial attention module to obtain a differential convolution feature matrix; The differential convolution feature matrix is input into a spatial attention part of the spatial attention module to obtain a spatial attention matrix; The spatial attention matrix is input into a Softmax activation function to obtain a spatial attention feature matrix; and The spatial attention feature matrix and the differential convolution feature matrix are multiplied by position points to obtain the classification feature matrix.

6. The method of claim 5, wherein the method is performed in a vacuum chamber. The classification feature matrix is input into a classifier to obtain a classification result, the classification result being used to represent whether the consistency between the security film pattern and the design pattern meets a predetermined standard, including: The classification feature matrix is unfolded into a classification feature vector in a row vector or a column vector; A plurality of fully connected layers of the classifier are used to perform fully connected encoding on the classification feature vector to obtain an encoded classification feature vector; and The encoded classification feature vector is input into a Softmax classification function of the classifier to obtain the classification result.

7. An inner series high sheet resistance film for a flexible HVDC capacitor, characterized by, The inner series high square resistance film is prepared by the preparation method of the inner series high square resistance film for flexible HVDC capacitor as claimed in any one of claims 1 to 6.

Images