초록 close

터빈 로타 디스크의 키웨이 근방에서 발생하는 균열을 검출하기 위해 초음파 신호처리 알고리즘을 개발하였다. 초음파 주사시스템을 이용한 B-주사 영상은 키웨이, 저면 그리고 균열에 대한 정보를 포함하고 있다. 개발한 알고리즘은 전처리, 균열후보영역의 검출, 균열영역의 판별의 3단계로 구성하였다. B-주사 영상에서 균열에 대한 초음파 영상을 구별해 내기 위해서, 균열이 존재하는 영상과 균열이 없는 영상을 비교하여 얻은 영상정보로부터 균열영역을 추출할 수 있었고, 전처리, 팽창처리, 영역분할 처리, 정합처리 그리고 차분처리과정을 통해 수행하였다. 개발한 초음파 신호처리 알고리즘을 이용해 터빈 로타 디스크 키웨이의 균열을 검출할 수 있음을 확인하였다.


An ultrasonic signal processing algorithm was developed to detect cracks generated around the keyway of a turbine rotor disk. The B-scan images obtained from the ultrasonic scanning system had ultrasonic images of the keyway, the backwall, and cracks. The developed algorithm was composed of three processing steps: pre-processing, detection of the candidate crack region, and classification of the crack region. The crack regions from the keyway and the backwall in the B-scan images could be detected from the obtained ultrasonic images by comparing them with the ultrasonic images of the keyway and the backwall with no cracks. The crack regions were detected by using binarization, dilation, segmentation, template matching, and subtraction. Experiments confirmed that the developed algorithm could detect the cracks.


An ultrasonic signal processing algorithm was developed to detect cracks generated around the keyway of a turbine rotor disk. The B-scan images obtained from the ultrasonic scanning system had ultrasonic images of the keyway, the backwall, and cracks. The developed algorithm was composed of three processing steps: pre-processing, detection of the candidate crack region, and classification of the crack region. The crack regions from the keyway and the backwall in the B-scan images could be detected from the obtained ultrasonic images by comparing them with the ultrasonic images of the keyway and the backwall with no cracks. The crack regions were detected by using binarization, dilation, segmentation, template matching, and subtraction. Experiments confirmed that the developed algorithm could detect the cracks.