Application of Machine Learning to Nuclear Safety Assessment - Professor Lee Jeong-ik (KAIST)

Machine learning is a favorable methodology when applied to physical problems that involve high levels of uncertainty or complex relationships among variables, making it difficult to express them directly through mathematical means. Accidents that could occur in nuclear power plants involve a multitude of interacting factors, including various equipment and operator actions that contribute causally, resulting in complex scenarios. Utilizing machine learning can help predict outcomes and derive optimal response strategies. In this webinar, we aim to introduce research conducted at KAIST, where machine learning is employed to predict accident outcomes and devise optimal response measures in scenarios similar to the Fukushima nuclear disaster.

[@MERRICS]

56:20 현재 원자력 발전소의 운영에는 현재 기계학습이 도입된 사례가 있는지?
57:10 원자력 발전소의 운영 중에 발생할 수 있는 문제나 안전 위험을
대량의 센서 데이터와 운영 기록을 기반으로 기계학습 모델을 훈련이 필요한데 연구에 필요한 실측데이트는 어떻게 구하는지?