This study proposes a methodology for predicting the physical properties such as the density of polymer composites, including asphalt binders, and evaluates its feasibility by identifying the quantitative relationship between the structure and properties of individual polymers. To this end, features are constructed using molecular dynamics (MD) simulation results and descriptor calculation tools. This study investigates the changes in the calculated density depending on the characteristics of the training dataset and analyzes the feature characteristics across datasets to identify key features. In this study, 2,415 hydrocarbon and binder-derived polymer molecules were analyzed using MD simulations and 2,790 chemical descriptors generated using alvaDesc. The features were pre-processed using correlation filtering, PCA, and recursive feature elimination. The XGBoost models were trained using k-fold cross-validation and Optuna optimization. SHAP analysis was used to interpret feature contributions. The variables influencing the density prediction differed between the hydrocarbon and binder groups. However, the hydrogen atom count (H), van der Waals energy, and descriptors such as SpMAD_EA_LboR consistently had a strong impact. The trained models achieved high accuracy (R² > 0.99) across different datasets, and the SHAP results revealed that the edge adjacency, topological, and 3D geometrical descriptors were critical. In terms of predictive accuracy and interpretability, the integrated MDQSPR framework demonstrated high reliability for estimating the properties of individual binder polymers. This approach contributed to a molecular-level understanding and facilitated the development of ecofriendly and efficient modifiers for asphalt binders.

This study proposes a methodology for predicting properties such as the density of polymer composites, including asphalt binders, and evaluates its feasibility by identifying the quantitative relationship between the structure and properties of individual polymers. To this end, this study investigates the variations in molecular dynamics (MD) results with molecular structural complexity and assesses the independence and correlation of variables that influence density. In this study, MD simulations were performed on hydrocarbon-based and individual asphalt binder molecules. The effects of various temperatures, molecular conditions, and structural features on the density were analyzed. MD-related variables influencing the calculated density were evaluated and compared with experimentally measured densities. The MD-calculated densities were used as target variables in a subsequent study, in which a machine learning model was applied to perform quantitative structure–property relationship analysis.The MD-calculated densities showed a strong correlation with experimental measurements, achieving a coefficient of determination of R2 > 0.95. Potential energy exhibited a tendency to cluster into 4–6 groups depending on the molecular structure. In addition, increasing molecular weight and decreasing temperature led to higher density and viscosity. Torsional energy and other individual energy components were identified as significant factors influencing both potential energy and density. This study provided foundational data for the property prediction of asphalt binders by quantitatively analyzing the relationship between the molecular structure and properties using MD simulations. Key features that could be used in the construction of polymer structure databases and AI-based material design were also proposed. In particular, the integration of MD-based simulation and machine learning was confirmed to be a practical alternative for predicting the properties of complex polymer composite systems.

This study determined the minimum size of a representative molecular structure for use in future dynamic analyses of asphalt binders. The minimum representative size, considering factors such as aging, additive types, and temperature variations, was established using density and radial distribution functions. This approach ensures that the structure reflects temperature-dependent property changes, which are critical characteristics of asphalt binders. In this study, the structure of asphalt-binder molecules was generated using the composition proposed by Li and Greenfield (2014) for AAA1. To assess the appropriateness of the molecular structure size, we generated additional structures, X2 and X3, maintaining the same composition as X1, but with two and three times the number of molecules, respectively, as suggested by Li and Greenfield (2014). Silica and lignin were considered as additives, and the aging conditions examined included unaged, short-term aging, and long-term aging. In addition, 11 temperature conditions were investigated. The density and radial distribution functions were plotted and analyzed. The variables influencing the density and radial distribution functions were set as the aging degree of the asphalt binder (unaged, short-term aging, long-term aging), 11 temperature conditions ranging from 233 to 433 K in 20 K intervals, structure size (X1, X2, and X3), and the presence of additives (no additives, silica, and lignin). For density, clear differences were observed based on the degree of aging, temperature conditions, and presence of additives, whereas the structure size did not significantly affect the density. In terms of radial distribution functions, the X1 structure reflected differences based on the degree of aging and the presence of additives but was limited in exhibiting temperature-dependent variations. In contrast, the X3 structure effectively captured temperature-dependent trends, indicating that the size of the molecular structure is crucial when evaluating energy calculations or physical tensile strength, necessitating careful assessment.

This paper explores a convergent approach that combines advanced informatics and computational science to develop road-paving materials. It also analyzes research trends that apply artificial-intelligence technologies to propose research directions for developing new materials and optimizing them for road pavements. This paper reviews various research trends in material design and development, including studies on materials and substances, quantitative structure–activity/property relationship (QSAR/QSPR) research, molecular data, and descriptors, and their applications in the fields of biomedicine, composite materials, and road-construction materials. Data representation is crucial for applying deep learning to construction-material data. Moreover, selecting significant variables for training is important, and the importance of these variables can be evaluated using Pearson’s correlation coefficients or ensemble techniques. In selecting training data and applying appropriate prediction models, the author intends to conduct future research on property prediction and apply string-based representations and generative adversarial networks (GANs). The convergence of artificial intelligence and computational science has enabled transformative changes in the field of material development, contributing significantly to enhancing the performance of road-paving materials. The future impacts of discovering new materials and optimizing research outcomes are highly anticipated.

음이온 교환막(AEM) 수전해용 AEM 소재 개발은 재생 에너지를 활용한 수소 생산 기술을 개선하는 데 중요한 역할을 한다. 이러한 소재를 설계하고 최적화하는 데 분자동역학 전산모사가 유용하게 사용되지만, 전산모사 결과의 정확도 는 사용된 force-field에 크게 의존한다. 본 연구의 목적은 AEM 소재의 구조와 이온 전도 특성을 예측할 때 force-field 선택 이 미치는 영향을 체계적으로 조사하는 것이다. 이를 위해 poly(spirobisindane-co-aryl terphenyl piperidinium) (PSTP) 구조를 모델 시스템으로 선택하고 COMPASS III, pcff, Universal, Dreiding 등 네 가지 주요 force-field를 비교 분석하였다. 각 force-field의 특성과 한계를 평가하기 위해 298~353 K의 온도 범위에서 수화 채널 형태, 물 분자와 수산화 이온의 분포, 수산 화 이온 전도성을 계산하였다. 이를 통해 AEM 소재의 분자동역학 전산모사에 가장 적합한 force-field를 제시하고, 고성능 AEM 소재 개발을 위한 계산 지침을 제공하고자 한다.

PURPOSES : As evaluation methods for road paving materials become increasingly complex, there is a need for a method that combines computational science and informatics for new material development. This study aimed to develop a rational methodology for applying molecular dynamics and AI-based material development techniques to the development of additives for asphalt mixtures. METHODS : This study reviewed relevant literature to analyze various molecular models, evaluation methods, and metrics for asphalt binders. It examined the molecular structures and conditions required for calculations using molecular dynamics and evaluated methods for assessing the interactions between additives and asphalt binders, as well as properties such as the density, viscosity, and glass transition temperature. Key evaluation indicators included the concept and application of interaction energy, work of adhesion, cohesive energy density, solubility parameters, radial distribution function, energy barriers, elastic modulus, viscosity, and stress-strain curves. RESULTS : The study identified key factors and conditions for effectively evaluating the physical properties of asphalt binders and additives. It proposed selective application methods and ranges for the layer structure, temperature conditions, and evaluation metrics, considering the actual conditions in which asphalt binders were used. Additional elements and conditions considered in the literature may be further explored, considering the computational demands. CONCLUSIONS : This study devised a methodology for evaluating the physical properties of asphalt binders considering temperature and aging. It reviewed and selected useful indicators for assessing the interaction between asphalt binders, additives, and modified asphalt binders and aggregates under various environmental conditions. By applying the proposed methods and linking the results with informatics, the interaction between asphalt binders and additives could be efficiently evaluated, serving as a reliable method for new material development.

디지털 트윈 기술의 도입은 소재/제품 개발 및 공정의 전주기 과정에서 보다 통합적이며 단절없는 디지털 가상화를 요구하고 있다. 이러한 요구는 미시적 반응, 표면 및 계면 현상을 아우르는 모델링 기법과 거시적 물리 모델 혹은 인공지능 모델의 광범위한 적용을 필요로 한다. 이는 다양한 환경조건에서 소재의 물성 데이터베이스와 미시적 현상 모사가 필요함을 의미하며, 분자동역학 시뮬레이션이 이를 달성하기 위하여 유용하게 활용될 수 있다. 본 논문에서는 평형 및 비평형 분자 동역학 시뮬레이션 방법을 활용한 물성 계산 방법을 개괄하고, 열 및 기계적 물성등 주요 물성 계산 사례들을 검토하여 제시하였다. 본 논문은 분자 동역학 시뮬레이션을 활용한 물성 계산 프레임 워크 개발과 보다 정확하며 신뢰도 높은 계산 수행에 통찰을 제공할 것으로 기대 된다.

컴퓨터 시스템의 성능 및 다양한 전산모사 프로그램의 발전으로 더 복잡한 원소로 이루어진 화학시스템의 해석이 가능해지고, 그에 따라 분자동역학 전사모사를 활용한 연구가 활발히 이루어지고 있다. 특히, 기존에는 실험위주로 진행되던 고분자 막에 대한 기체 투과 특성을 계산하는 연구가 관심을 받고 있고, 식품포장, 의약품등에 사용되고 있는 기체차단성 막 에 대한 분자동역학 연구가 많이 이루어지고 있다. 최근 실크 피브로인을 이용해 코팅막을 만들었을 때 기체 차단 효과가 나 타난다는 보고가 있었고, 본 연구에서는 이러한 실크 피브로인을 활용해 복합막을 만들었을 때 산소 차단 효과가 나타나는지 확인하고자 분자동역학 전산모사를 이용해 연구를 진행하였다. 단일 모델을 제작하고 기체 투과 특성을 계산하고 실험값과 비교를 통해 모델이 실제 실험 결과를 반영하는 것을 확인하였고, 실제 복합막 모델을 만들어 고분자 내에서 기체 이동경로 분석을 진행한 결과 산소 분자가 피브로인 영역을 통과하지 못하고 막히는 것을 보여주었다. 따라서, 실크 피브로인이 도입된 복합막이 산소 차단 성능이 우수하여, 식품포장 등에 유용할 것으로 기대된다.

PURPOSES : In this study, we aimed to evaluate the transition temperature (Tt) of asphalt binders using molecular dynamics simulations, which can provide a more accurate assessment of the mechanical properties of a material at the molecular level and can be applied to material development and design. METHODS : Unlike conventional macro- or meso-level simulations, we utilized MD simulations to evaluate the Tg of asphalt binders based on material composition and aging degree as input variables. In this analysis, 11 temperatures ranging from 434 K to 233 K at 20 K intervals were utilized, and the bulk volume and density were calculated through MD simulations. RESULTS : The MD simulation successfully predicted the Tg of the asphalt binder, and the molecular-level properties and interactions determined in this study can be applied not only to material development but also to the determination of constitutive equations or contact models used in continuum mechanics or discrete element methods. The calculated Tg was slightly different depending on the aging of the asphalt binder; however, it was found to accurately reflect the transitional characteristics. CONCLUSIONS : This study demonstrated the potential of MD simulations as valuable tools for material development and design in the construction industry. The results indicate that the use of MD simulations can lead to more accurate and efficient material development and design by providing a more detailed understanding of material properties and interactions at the molecular level.

금속의 취성화는 수소와 접촉하는 구조물을 안정적으로 설계하는데 있어서 큰 문제가 되어왔다. 본 논문에서는 분자동역학 해석을 통해 균열선단 주변에 모인 수소원자들이 전위 이동 현상을 억제하고, 이로 인해 벽개 파괴 현상이 발생하는 것을 확인하였다. 다양한 수소 농도, 하중 속도, 수소 확산 속도 등을 바꾸어가며 분자동역학 해석을 수행하였고, 이에 따른 수소 취성화를 최소화시킬 수 있는 조건들을 조사하였다. 분자동역학 해석 결과는 기존의 실험결과와 잘 일치하였으며 이를 바탕으로 수소 취성화 현상을 정량화하여 평가하였다.