http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
자기애자의 유지 관리를 위한 CNN 기법을 이용한 이미지 분석
최인혁,신구용,구자빈,손주암,임대연,오태근,윤영근,Choi, In-Hyuk,Shin, Koo-Yong,Koo, Ja-Bin,Son, Ju-Am,Lim, Dae-Yeon,Oh, Tae-Keun,Yoon, Young-Geun 한국전기전자재료학회 2020 전기전자재료학회논문지 Vol.33 No.3
This study examines the feasibility of the image deep learning method using convolution neural networks (CNNs) to maintain a porcelain insulator. Data augmentation is performed to prevent over-fitting, and the classification performance is evaluated by training the age, material, region, and pollution level of the insulator using image data in which the background and labelling are removed. Based on the results, it was difficult to predict the age, but it was possible to classify 76% of the materials, 60% of the pollution level, and more than 90% of the regions. From the results of this study, we identified the potential and limitations of the CNN classification for the four groups currently classified. However, it was possible to detect discoloration of the porcelain insulator resulting from physical, chemical, and climatic factors. Based on this, it will be possible to estimate the corrosion of the cap and discoloration of the porcelain caused by environmental deterioration, abnormal voltage, and lightning.
최인혁,신구용,임윤석,구자빈,손주암,임대연,오태근,윤영근,Choi, In-Hyuk,Shin, Koo-Yong,Lim, Yun-seog,Koo, Ja-Bin,Son, Ju-Am,Lim, Dae-Yeon,Oh, Tae-Keun,Yoon, Young-Geun 한국전기전자재료학회 2020 전기전자재료학회논문지 Vol.33 No.3
This paper investigates the soundness of porcelain insulators associated with the acoustic emission (AE) technique. The AE technique is a popular non-destructive method that measures and analyzes the burst energy that occurs mainly when a crack occurs in a high-frequency region. Typical AE methods require continuous monitoring with frequent sensor calibration. However, in this study, the AE technique excites a porcelain insulator using only an impact hammer, and it applies a high-pass filter to the signal frequency range measured only in the AE sensor by comparing the AE and the acceleration sensors. Next, the extracted time-domain signal is analyzed for the damage assessment. In normal signals, the duration is about 2ms, the area of the envelope is about 1,000, and the number of counts is about 20. In the damage signal, the duration exceeds 5ms, the area of the envelope is about 2,000, and the number of counts exceeds 40. In addition, various characteristics in the time and frequency domain for normal and damage cases are analyzed using the short-time Fourier transform (STFT). Based on the results of the STFT analysis, the maximum energy of a normal specimen is less than 0.02, while in the case of the damage specimen, it exceeds 0.02. The extracted high-frequency components can present dynamic behavior of crack regions and eigenmodes of the isolated insulator parts, but the presence, size, and distribution of cracks can be predicted indirectly. In this regard, the characteristics of the surface crack region were derived in this study.