A method for checking and identifying the polarity and ratio markings of distribution network instrument transformers

Abstract

This application relates to a method for checking and identifying the polarity and ratio markings of distribution network transformers. The method includes establishing a database of the appearance structure and ratio and polarity markings of distribution network transformers from various manufacturers, models, and specifications; collecting images of these transformers from various angles and training the database using the YOLO algorithm; capturing images of the transformer ratio and polarity markings using a camera; using a deep learning-based object detection algorithm to classify the transformers in the images; extracting the ratio and primary / secondary polarity marking positions of similar distribution network transformers from the database; and again using a deep learning-based object detection algorithm to locate the polarity and ratio markings in the input images and extracting the location information. The method identifies the primary terminal polarity marking, secondary terminal polarity marking, and ratio marking in the located images and determines whether the markings and ratio markings of the primary and secondary terminals are correct. This application overcomes the errors of manual operation and greatly improves work efficiency.

Description

Technical Field

[0001] This application relates to the field of distribution network instrument transformer inspection, and in particular to a method for inspecting and identifying the polarity and ratio markings of distribution network instrument transformers. Background Technology

[0002] Checking the transformer ratio and polarity markings is a necessary step in transformer testing, affecting the accuracy of information and the safety of the test. Currently, it is generally done manually, but this method is prone to human error and has low efficiency. Summary of the Invention

[0003] The purpose of this application is to provide a method for checking and identifying the polarity and ratio markings of distribution network transformers, which can overcome the errors of manual operation and greatly improve work efficiency.

[0004] To achieve the above objectives, this application provides the following technical solution:

[0005] This application provides a method for checking and identifying the polarity and ratio markings of distribution network instrument transformers, including the following specific steps:

[0006] Establish a database of the appearance structure, transformation ratio and polarity identification of distribution network transformers of various manufacturers, models and specifications, and collect pictures of these transformers from various angles, and use the YOLO algorithm for training.

[0007] Use cameras to capture images of the transformer ratio and polarity markings of power distribution network transformers;

[0008] Using a deep learning-based object detection algorithm, the transformers in the image are identified and classified. Then, the transformation ratio and primary and secondary polarity markings of similar distribution network transformers are extracted from the database. The deep learning-based object detection algorithm is used again to locate the polarity and transformation ratio markings in the input image and extract the location information. The primary polarity marking, secondary polarity marking, and transformation ratio markings of the target are located in the image, and the correctness of the markings of the primary and secondary terminals and the transformation ratio markings is determined.

[0009] The training using the YOLO algorithm specifically involves determining the bounding box and anchor box dimensions of the detection target based on the collected images of the transformer from various angles, combined with the appearance structure of the distribution network transformer and the database of transformation ratio and polarity identifiers, and obtaining the prediction model.

[0010] The method of using a deep learning-based target detection algorithm to identify and classify transformers in images involves inputting images of transformer ratios and polarity markings captured by a camera into a prediction model trained with the YOLO algorithm. The prediction model then marks anchor frames for the ratios and polarities, predicts whether the anchor frame regions contain the ratios and polarity markings, and identifies whether a transformer is a voltage transformer or a current transformer based on the prediction results.

[0011] The next step involves extracting the transformation ratio and primary / secondary polarity identifier positions of similar distribution network transformers from the database. Then, a deep learning target detection algorithm is used again to locate the polarity and transformation ratio identifiers in the input image and extract the location information. Specifically, the primary polarity identifier, secondary polarity identifier, and transformation ratio identifier in the target location image are determined by using the appearance structure of the distribution network transformers and the transformation ratio and polarity identifier database. Based on the location of the transformation ratio and primary / secondary polarity identifiers, the anchor frame size and position of the prediction model are adjusted. The adjusted anchor frame size and position of the prediction model are then matched and predicted with the identifiers in the image to obtain the transformation ratio and primary / secondary polarity identifiers of the transformers.

[0012] Compared with existing technologies, the beneficial effects of this application are as follows: This application utilizes a visual sensor to replace the human eye, acquiring images of the transformer at the detection station and converting them into a data matrix. It uses a computer to replace the human brain, employing software for image processing and analysis, thereby achieving unified collection and management of equipment information and improving the accuracy of determining the transformer ratio and polarity indicators. This overcomes the errors inherent in manual operation and significantly improves work efficiency. Attached Figure Description

[0013] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a schematic diagram of the method flow of the present invention.

[0015] Figure 2 A photograph of a current transformer;

[0016] Figure 3 Anchor frame annotation of the physical diagram of the current transformer ratio parameters identified by the method of this application;

[0017] Figure 4 Anchor frame annotation of a physical diagram of the current transformer polarity parameters identified by the method of this application;

[0018] Figure 5 A photograph of a voltage transformer;

[0019] Figure 6 Anchor frame annotation of physical diagram of voltage transformer level parameters identified by the method of this application;

[0020] Figure 7 The anchor frame annotation of the physical diagram of the voltage transformer polarity parameters identified by the method of this application; Detailed Implementation

[0021] The technical solutions of the embodiments of this application will now be described with reference to the accompanying drawings. It should be noted that similar reference numerals and letters in the following drawings indicate similar items; therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0022] The terms “comprising,” “including,” or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase “comprising one…” does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0023] like Figure 1 As shown, a method for checking and identifying the polarity and ratio markings of distribution network instrument transformers includes the following specific steps:

[0024] Establish a database of the appearance structure, transformation ratio and polarity identification of distribution network transformers of various manufacturers, models and specifications, and collect pictures of these transformers from various angles, and use the YOLO algorithm for training.

[0025] Use cameras to capture images of the transformer ratio and polarity markings of power distribution network transformers;

[0026] Using a deep learning-based object detection algorithm, the transformers in the image are identified and classified. Then, the transformation ratio and primary and secondary polarity markings of similar distribution network transformers are extracted from the database. The deep learning-based object detection algorithm is used again to locate the polarity and transformation ratio markings in the input image and extract the location information. The primary polarity marking, secondary polarity marking, and transformation ratio markings of the target are located in the image, and the correctness of the markings of the primary and secondary terminals and the transformation ratio markings is determined.

[0027] The training using the YOLO algorithm specifically involves determining the bounding box and anchor box dimensions of the detection target based on the collected images of the transformer from various angles, combined with the appearance structure of the distribution network transformer and the database of transformation ratio and polarity identifiers, and obtaining the prediction model.

[0028] The method of using a deep learning-based target detection algorithm to identify and classify transformers in images involves inputting images of transformer ratios and polarity markings captured by a camera into a prediction model trained with the YOLO algorithm. The prediction model then marks anchor frames for the ratios and polarities, predicts whether the anchor frame regions contain the ratios and polarity markings, and identifies whether a transformer is a voltage transformer or a current transformer based on the prediction results.

[0029] The next step involves extracting the transformation ratio and primary / secondary polarity identifier positions of similar distribution network transformers from the database. Then, a deep learning target detection algorithm is used again to locate the polarity and transformation ratio identifiers in the input image and extract the location information. Specifically, the primary polarity identifier, secondary polarity identifier, and transformation ratio identifier in the target location image are determined by using the appearance structure of the distribution network transformers and the transformation ratio and polarity identifier database. Based on the location of the transformation ratio and primary / secondary polarity identifiers, the anchor frame size and position of the prediction model are adjusted. The adjusted anchor frame size and position of the prediction model are then matched and predicted with the identifiers in the image to obtain the transformation ratio and primary / secondary polarity identifiers of the transformers.

[0030] like Figures 2-4 As shown, Figure 2 A photograph of a current transformer is input into the model trained in this application, which can then identify features such as... Figure 3 and Figure 4 The transformer ratio parameters and primary and secondary polarities are shown in the figure. P1 and P2 are the primary polar and non-polar terminals, and s1 and s2 are the secondary polar and non-polar terminals.

[0031] Figures 5-7 A photograph of a voltage transformer is input into the model trained in this application, which can identify features such as... Figure 6 and Figure 7 The transformer ratio parameters and primary and secondary polarities are shown in the diagram. In the diagram, A and B are the primary polarity terminals and the non-polarity terminals, and a and b are the secondary polarity terminals and the non-polarity terminals.

[0032] Using the method described in this application, the applicant tested 516 current transformers and 472 voltage transformers that had been purchased. The overall identification accuracy was 83.72% for current transformers and 85.17% for voltage transformers. The overall identification accuracy is relatively high, but further optimization is still needed.

[0033] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for checking and identifying the polarity and ratio markings of distribution network instrument transformers, characterized in that, The specific steps include the following: Establish a database of the appearance structure, transformation ratio and polarity identification of distribution network transformers of various manufacturers, models and specifications, and collect images of these transformers from various angles, and use the YOLO algorithm for training; Use cameras to capture images of the transformer ratio and polarity markings of power distribution network transformers; Using a deep learning-based object detection algorithm, the transformers in the image are identified and classified. Then, the transformation ratio and primary and secondary polarity labels of similar distribution network transformers are extracted from the database. Again, the deep learning-based object detection algorithm is used to locate the polarity and transformation ratio labels in the input image and extract the location information. The primary polarity label, secondary polarity label, and transformation ratio label are located in the image with the target, and the correctness of the primary and secondary terminal labels and transformation ratio labels is determined. The training using the YOLO algorithm specifically involves determining the bounding box and anchor box dimensions of the detection target based on the collected images of the transformer from various angles, combined with the appearance structure of the distribution network transformer and the database of transformation ratio and polarity identifiers, and obtaining the prediction model. The method of using a deep learning-based target detection algorithm to identify and classify transformers in images specifically involves inputting images of transformer ratios and polarity markings captured by a camera into a prediction model trained with the YOLO algorithm. The prediction model marks anchor frames for the ratios and polarities, predicts whether the anchor frame area contains the ratios and polarity markings, and identifies whether it is a voltage transformer or a current transformer based on the prediction results. The next step involves extracting the transformation ratio and primary / secondary polarity identifier positions of similar distribution network transformers from the database. Then, a deep learning target detection algorithm is used again to locate the polarity and transformation ratio identifiers in the input image and extract the location information. Specifically, the primary polarity identifier, secondary polarity identifier, and transformation ratio identifier in the target location image are determined by using the appearance structure of the distribution network transformers and the transformation ratio and polarity identifier database. Based on the location of the transformation ratio and primary / secondary polarity identifiers, the anchor frame size and position of the prediction model are adjusted. The adjusted anchor frame size and position of the prediction model are then matched and predicted with the identifiers in the image to obtain the transformation ratio and primary / secondary polarity identifiers of the transformers.

Images