http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Jeong, Gil-Eon,Choi, Woo-Seok,Cho, Sang Soon Korean Nuclear Society 2021 Nuclear Engineering and Technology Vol.53 No.7
Spent nuclear fuel, which can degrade during long-term storage, must be transported intact in normal transport conditions. In this regard, many studies, including those involving Multi-Modal Transportation Test (MMTT) campaigns, have been conducted. In order to transport the spent fuel safely, a tie-down structure for supporting and transporting a cask containing the spent fuel is essential. To ensure its structural integrity, a method for finding an optimum conceptual design for the tie-down structure is presented. An optimized transportation test model of a tie-down structure for the KORAD-21 metal cask is derived based on the proposed optimization approach, and the transportation test model is manufactured by redesigning the optimized model to enable its producibility. The topology optimization approach presented in this paper can be used to obtain optimum conceptual designs of tie-down structures developed in the future.
Training Data Production for Developing Machine Learning Based Hybrid Solver for Disposal Repository
Gil-Eon Jeong,DongHyuk Lee,Hong Jang,Jung-Woo Kim 한국방사성폐기물학회 2023 한국방사성폐기물학회 학술논문요약집 Vol.21 No.1
To conduct numerical simulation of a disposal repository of the spent nuclear fuel, it is necessary to numerically simulate the entire domain, which is composed on numerous finite elements, for at least several tens of thousands of years. This approach presents a significant computational challenge, as obtaining solutions through the numerical simulation for entire domain is not a straightforward task. To overcome this challenge, this study presents the process of producing the training data set required for developing the machine learning based hybrid solver. The hybrid solver is designed to correct results of the numerical simulation composed of coarse elements to the finer elements which derive more accurate and precise results. When the machine learning based hybrid solver is used, it is expected to have a computational efficiency more than 10 times higher than the numerical simulation composed of fine elements with similar accuracy. This study aims to investigate the usefulness of generating the training data set required for the development of the hybrid solver for disposal repository. The development of the hybrid solver will provide a more efficient and effective approach for analyzing disposal repository, which will be of great importance for ensuring the safe and effective disposal of the spent nuclear fuel.
Gil-Eon Jeong,Dong-Won Jang 한국방사성폐기물학회 2022 한국방사성폐기물학회 학술논문요약집 Vol.20 No.1
The amount of temporarily stored spent nuclear fuel in South Korea will be reaching saturation in a near future. Therefore, it is an urgent issue to construct a spent nuclear fuel storage system. In order to construct the storage system, some coastal environmental characteristics such as temperature, pH, and chemical composition of sea water in South Korea have to be evaluated and predicted because they can affect in deterioration of the storage system. However, in South Korea, the coastal environmental characteristics of area where the storage system is likely to be built are not well established until now. In this study, a time-series deep-learning algorithm is developed using the Long-Short Term Memory (LSTM) algorithm to predict and evaluate the coastal environmental characteristics based on the wellestablished data from Korea Meteorological Administration (KMA) and Ministry of Oceans and Fisheries (MOF). As a result, by developing the predictive model to evaluate the coastal environmental characteristics, we intend to apply it for site evaluation to construct the spent nuclear fuel storage system or many other applications related to the nuclear as well.
Surrogate Modeling of Disposal System for Machine Learning Based Hybrid Solver Using U-Network
Gil-Eon Jeong,DongHyuk Lee,Hong Jang,Jung-Woo Kim 한국방사성폐기물학회 2023 한국방사성폐기물학회 학술논문요약집 Vol.21 No.2
Conducting a TSPA (Total System Performance Assessment) of the entire spent nuclear fuel disposal system, which includes thousands of disposal holes and their geological surroundings over many thousands of years, is a challenging task. Typically, the TSPA relies on significant efforts involving numerous parts and finite elements, making it computationally demanding. To streamline this process and enhance efficiency, our study introduces a surrogate model built upon the widely recognized U-network machine learning framework. This surrogate model serves as a bridge, correcting the results from a detailed numerical model with a large number of small-sized elements into a simplified one with fewer and large-sized elements. This approach will significantly cut down on computation time while preserving accuracy comparable to those achieved through the detailed numerical model.