Sungyoung Kim
Korea Advanced Institute of Science and Technology · 工学
研究室紹介
Professor Sungyoung Kim's research lab specializes in smart sensing and structural health monitoring, focusing on non-destructive evaluation techniques for critical infrastructure and mechanical systems. The lab develops advanced magnetic and electromagnetic sensing technologies—such as magnetic flux leakage (MFL) and elasto-magnetic sensors—for real-time detection of damage in components like flat belts, tension members, and civil infrastructure. By integrating machine learning models like GRU networks with sensor data, the lab enables high-accuracy condition monitoring and predictive maintenance. The research also extends to image-based crack detection in civil structures using deep learning, enhancing the reliability and efficiency of infrastructure management.
Research Overview
Research Output Trend
Figures are computed from collected data and may differ slightly.
Selected Papers
8Flat belts are increasingly used in elevators, which offer faster stabilization and energy savings compared to wire ropes. How- ever, damage to flat belts during operation can lead to catastrophic accidents, such as rope failure and falls due to tensile loads. Therefore, there is a need for monitoring techniques to detect damage in advance and prevent accidents. Although extensive research has been conducted on the diagnosis of damage to wire ropes, studies on diagnosing damage to flat belts are
This letter proposes a method to the estimation of tension force in tension members using the grated recurrent unit (GRU) algorithm. In this letter, a yoke-type elasto-magnetic (E/M) sensor was developed based on numerical ANSYS Maxwell simulations to enhance the applicability through the structural improvement of the existing solenoid-type magnetized E/M sensor. The induced voltage signal collected based on the yoke-type E/M sensor was applied to the GRU algorithm. As a result of applying the G
포장도로와 콘크리트 표면의 균열 유무를 정확히 식별하는 것은 인프라 안전 및 관리의 핵심적인 부분이며, 최근에는 자동화된 영상 분석 기반의 균열 탐지 기술이 유지관리 효율성을 향상시키는 핵심 도구로 주목받고 있다. 본 논문에서는 DRAEM 모델을 개선하여 균열 감지의 효율성과 정확도를 높이기 위한 새로운 방법을 제시한다. 제안하는 방법은 균열의 효과적인 증강, 노이즈 감소를 위한 필터링 및 모델 구조의 최적화에 중점을 두어, 기존 모델보다 우수한 성능을 달성하는 것을 목표로 한다. 여러 지표를 통해 성능을 비교한 결과, 특히 AUC 지표에서 약 23%의 성능 향상이 있음을 확인하였다. 이러한 기술은 인프라의 안전성을 향상시키고 유지관리 비용을 줄이는 데 결정적인 역할을 할 것으로 기대된다.