Search

KR-102963423-B1 - DETERIORATION DEGREE EVALUATION APPARATUS AND METHOD FOR EVALUATING DETERIORATION DEGREE

KR102963423B1KR 102963423 B1KR102963423 B1KR 102963423B1KR-102963423-B1

Abstract

A method for evaluating the degree of deterioration according to one embodiment of the present invention comprises the steps of: photographing a plurality of sections of a certain facility to generate a plurality of image data; binaryizing the plurality of image data to generate a plurality of binary images; extracting a plurality of piping information included in the plurality of binary images; inputting the plurality of piping information into a pre-trained algorithm to group the plurality of binary images and sequence the degree of deterioration of each group; and determining the final number of groups and the degree of deterioration of each group based on at least one of the operation history and maintenance information of the facility.

Inventors

  • 이한상
  • 김범신
  • 김진수
  • 김희선
  • 박명수
  • 이창민

Assignees

  • 한국전력공사
  • 한국남동발전 주식회사
  • 한국중부발전(주)
  • 한국서부발전 주식회사
  • 한국남부발전 주식회사
  • 한국동서발전(주)

Dates

Publication Date
20260513
Application Date
20210527

Claims (20)

  1. A step of capturing multiple sections of a certain facility to generate multiple image data; A step of generating a plurality of binary images by binaryizing the above plurality of image data; A step of extracting multiple pipe information included in the above multiple binary images; A step of inputting the plurality of pipe information into a pre-learned algorithm to group the plurality of binary images and sequence the degree of degradation of each group; and Based on at least one of the operation history and maintenance information of the above-mentioned equipment, the method includes the step of determining the final number of groups and the degree of deterioration of each group. The step of extracting the above plurality of pipe information is, A method for evaluating the degree of deterioration characterized by extracting the plurality of pipe information using a contour technique that indicates the boundaries of particles.
  2. In Article 1, The step of generating the above plurality of binary images is, A degradation evaluation method characterized by generating a plurality of binary images from the plurality of image data using an adaptive threshold technique.
  3. In Article 2, The step of generating the above plurality of binary images is, A method for evaluating degradation, characterized by dividing the image into multiple regions for the above-mentioned multiple image data, setting a threshold value, and dividing the image into two regions by setting the contrast ratio of the threshold value as a boundary to generate the above-mentioned multiple binary images.
  4. In Article 1, A method for evaluating the degree of deterioration, characterized in that the plurality of pipe information above includes at least one of particle size, number, and circumference.
  5. delete
  6. In Article 1, A method for evaluating the degree of degradation, characterized in that the above contour technique is a technique for identifying particle boundaries by connecting the edge boundaries of parts having the same color or the same color intensity.
  7. In Article 1, The aforementioned previously learned algorithm includes a clustering technique, which is one of the unsupervised learning methods, and The step of ranking the degree of degradation of each of the above groups is, A degradation evaluation method characterized by grouping a plurality of binary images using a clustering technique, which is one of the above unsupervised learning methods.
  8. In Article 7, A method for evaluating the degree of degradation characterized in that the above clustering technique is a merged clustering technique.
  9. In Article 1, The step of ranking the degree of degradation of each of the above groups is, A degradation evaluation method characterized by using a dendrogram to group multiple binary images into an arbitrary number of groups and classifying the degradation level of each group.
  10. In Article 1, A method for evaluating the degree of degradation that further includes the step of generating a damage map by linking images classified by each group.
  11. An image acquisition module that captures multiple sections of a certain facility and generates multiple image data; and A control unit comprising: binaryizing the plurality of image data to generate a plurality of binary images; extracting a plurality of piping information included in the plurality of binary images; inputting the plurality of piping information into a pre-learned algorithm to group the plurality of binary images, sequence the degree of deterioration of each group, and determining the final number of groups and the degree of deterioration of each group based on at least one of the operation history and maintenance information of the equipment. The above control unit is, A deterioration evaluation device characterized by extracting the plurality of pipe information using a contour technique that indicates the boundaries of particles.
  12. In Article 11, The above control unit is, A degradation evaluation device characterized by generating a plurality of binary images from the plurality of image data using an adaptive threshold technique.
  13. In Article 12, The above control unit is, A degradation evaluation device characterized by, for the above-mentioned plurality of image data, dividing the image into multiple regions, setting a threshold value, and dividing it into two regions by setting the contrast ratio of the threshold value as a boundary to generate the above-mentioned plurality of binary images.
  14. In Article 11, A deterioration evaluation device characterized in that the above plurality of pipe information includes at least one of particle size, number, and circumference.
  15. delete
  16. In Article 11, A degradation evaluation device characterized by the above contour technique being a technique for identifying particle boundaries by connecting the edge boundaries of parts having the same color or the same color intensity.
  17. In Article 11, The aforementioned previously learned algorithm includes a clustering technique, which is one of the unsupervised learning methods, and The above control unit is, A degradation evaluation device characterized by grouping a plurality of binary images using a clustering technique, which is one of the above unsupervised learning methods.
  18. In Article 17, A degradation evaluation device characterized by the above clustering technique being a merged clustering technique.
  19. In Article 11, The above control unit is, A degradation evaluation device characterized by using a dendrogram to group a plurality of binary images into an arbitrary number of groups and classifying the degradation level of each group.
  20. In Article 11, The above control unit is, A degradation evaluation device characterized by generating a damage map by connecting images classified by each group.

Description

Degradation Degree Evaluation Apparatus and Method for Evaluating Degradation Degree The present invention relates to a deterioration evaluation device and a deterioration evaluation method, and more specifically, to a device and method capable of evaluating the degree of deterioration by utilizing image preprocessing and unsupervised learning classification techniques for equipment defect detection. When performing preventive maintenance on major equipment such as boilers and turbines in power plants, a technology is used to evaluate the type of damage and the grade of degradation by non-destructively evaluating the degree of degradation of the target parts through a grinding and etching process to replicate the metal surface as shown in Fig. 1 and then comparing it with a reference image as shown in Fig. 2. Figure 1 is a drawing showing a surface replica and a film after replication for boiler damage assessment. Figure 2 shows the standard images of X20 steel, a representative steel grade applied to standard coal-fired power plants, according to the VTT standard deterioration grade. The microstructure is divided into six grades from A to F based on the heat treatment temperature and time, and the final life consumption rate is calculated by linking this with the creep pore and microcrack formation criteria. Conventional methods for evaluating the degree of degradation have the following problems. First, the images used for classifying degradation in existing VTT reports were created by exposing test specimens of each steel grade to a constant temperature for a set period of time, which differs significantly from the actual operating environment of power plant facilities. During equipment operation, the temperature of tubes and pipes is not maintained at a constant level, and even the same component experiences temperature variations depending on its location. In addition, it is exposed to pressure caused by internal fluids, thermal stress due to temperature changes and contraction/expansion, residual stress generated during the fabrication of pressure vessels, and various multi-axial stress states depending on the component shape. Therefore, rather than the existing method of fabricating test specimens based solely on temperature and classifying the degree of degradation based on them, determining the order and classifying grades by relatively comparing images of actual parts is a method that can increase reliability. Secondly, while the comparison and classification of existing surface replica images are performed by experts in the field, human error may occur as the amount of data acquired in the field increases and interpretation takes a long time. In addition, since power generation facilities are assessed by different experts over an extended period, a problem may arise where grading standards vary from person to person. Figure 3 compares some images from the existing analysis results. The two figures are the results of analyzing different facilities in the same year. The left figure is an image of Unit 4, which was determined to be at stage E. The right figure is an image of Unit 6, which was determined to be at stage DE, which is intermediate between stages D and E. When comparing the two, it was determined that the E stage on the left had progressed relatively more than the DE stage on the right. However, when compared once again with the standard images for each grade in Figure 2, which serve as the basis for judgment, it can be seen that the right image has progressed more than the left image. Therefore, it can be seen that errors can frequently occur during the process of experts evaluating images. Finally, existing VTT reports provide only one or two photos per grade, making it difficult to determine the accurate grade. Accurate classification is possible only by providing various images according to grain size and magnification, and it is also difficult to use them as a reference point for image classification due to low resolution and contrast ratio. In other words, conventional degradation evaluation methods can lead to errors during the process of repeatedly analyzing large amounts of data when determining image degradation grades based on the judgment of existing experts. Additionally, errors may occur because standards differ due to multiple experts making judgments over a long period of equipment operation. It is necessary to develop an algorithm that prevents human error and enables rapid diagnosis by applying artificial intelligence techniques to classify grades while minimizing judgment based on human intuition during the analysis of such metal tissue replica images. Figures 1, 2, and 3 are conceptual diagrams for explaining a conventional method for evaluating the degree of degradation. FIG. 4 is a conceptual diagram showing the degradation evaluation device of the present invention. FIG. 5 is a flowchart illustrating a representative method for evaluating the degree of degradation