http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
데이터 최적화 기법을 도입한 기계적 물성치 예측 모델과 새로운 2차원 소재의 발견
이인효(Inhyo Lee),김준철(Joonchul Kim),김태현(Taehyun Kim),민경민(Kyoungmin Min) 대한기계학회 2022 대한기계학회 춘추학술대회 Vol.2022 No.11
Two-dimensional (2D) materials are attractive materials. Many studies are being conducted because of their unique characteristics. However, there is lack of information about properties of 2D materials. Therefore, this study attempted to solve this problem by developing a machine learning (ML) model that predicts mechanical properties of 2D materials. In addition, a 2D materials generation framework was developed using a classification model and a deep learning-based generative model. ML model to predict mechanical properties is trained from existing 2D database and reduces the uncertainty of prediction through data optimization techniques. Potential 2D materials are discovered through screening processes such as measuring structure and atomic similarities. We believe that the developing of ML model and framework for finding new 2D materials could open a new chapter in material science
데이터 최적화 기법을 도입한 기계적 물성치 예측 모델과 새로운 2차원 소재의 발견
이인효(Inhyo Lee),김준철(Joonchul Kim),김태현(Taehyun Kim),민경민(Kyoungmin Min) 대한기계학회 2022 대한기계학회 춘추학술대회 Vol.2022 No.11
Two-dimensional (2D) materials are attractive materials. Many studies are being conducted because of their unique characteristics. However, there is lack of information about properties of 2D materials. Therefore, this study attempted to solve this problem by developing a machine learning (ML) model that predicts mechanical properties of 2D materials. In addition, a 2D materials generation framework was developed using a classification model and a deep learning-based generative model. ML model to predict mechanical properties is trained from existing 2D database and reduces the uncertainty of prediction through data optimization techniques. Potential 2D materials are discovered through screening processes such as measuring structure and atomic similarities. We believe that the developing of ML model and framework for finding new 2D materials could open a new chapter in material science
반데르발스-헤테로 구조 데이터 베이스 구축 및 기계학습 기반 원자간 포텐셜 모델 개발
이인효(Inhyo Lee),김주오(Juo Kim),민경민(Kyoungmin Min) 대한기계학회 2023 대한기계학회 춘추학술대회 Vol.2023 No.11
Van der Waals (VdW) heterostructures hold immense potential as materials with versatile applications. Considerable research is underway to investigate their unique electronic and structural properties. However, conventional trial-anderror methods and density functional theory (DFT) calculations are inefficient in exploring the extensive materials space of VdW-heterostructures. Thus, this study aims to develop a machine learning interatomic potentials model for both Twodimensional (2D) materials structures and VdW heterostructure. To accomplish this goal, new VdW heterostructure data was created using the previously reported Two-dimensional materials database and DFT calculation. The development of MLIPs and the constructing of a heterostructure database can be widely used in the materials study field.
차세대 칼슘 이온 배터리 양극재 설계 및 선별을 위한 기계학습 플랫폼
김민선(Minseon Kim),박재정(Jaejung Park),김희규(Heekyu Kim),이재준(Jaejun Lee),이인효(Inhyo Lee),이승철(Seungchul Lee),민경민(Kyoungmin Min) 대한기계학회 2023 대한기계학회 춘추학술대회 Vol.2023 No.11
Machine learning was generated for developing highly stable cathode materials with the Ca-Ion Battery NASICON structure. The database is divided into a training set of 146,309 materials and a test set of 630 materials with newly designed NASICON structures. Employing 149 descriptors, including 147 chemical features and 2 structural features derived from the composition of each material. Random forest (RF) regressor, employed for Eform prediction, demonstrated impressive results with an R-squared of 0.916, MAE of 0.142, and RMSE of 0.351 eV/atom. Similarly, the RF classifier used for Ehull prediction exhibited an Accuracy of 0.818, AUC of 0.889, and Precision of 0.826. The optimal model was subsequently applied to predict stable materials among the 630 materials, based on the criteria of (1) Eform < 0 eV/atom and (2) Ehull < 0.05 eV/atom. As a result, 125 materials were identified as possessing both structural and thermodynamic stability in charge and discharge states.