Enhancing Phase Change Thermal Energy Storage Material Properties Prediction with Digital Technologies

Minghao Yu^1*Jing Liu²Cheng Chen³Mingyue Li¹

• ¹Dalian Maritime University, Dalian, China
• ²Dalian Ocean University, Dalian, Liaoning, China
• ³China National Offshore Oil Corporation (China) Limited Tianjin Branch, Tianjin, China

In the field of materials science, the prediction of material properties plays a critical role in designing new materials and optimizing existing ones. Traditional experimental approaches, while effective, are resource-intensive and time-consuming, often requiring extensive trial-anderror methods. To address these limitations, the integration of digital technologies, such as computational modeling and machine learning (ML), has become increasingly important. This paper proposes a hybrid multiscale modeling framework that integrates molecular dynamics (MD) simulations, finite element methods (FEM) from continuum mechanics, and supervised ML algorithms-including deep neural networks and gradient boosting regressors-to enable accurate and efficient prediction of material properties across scales. Traditional models often struggle with bridging the gap between scales or fail to incorporate large datasets for predictive accuracy and generalizability. In contrast, our method integrates MD simulations for atomic-level interactions using Lennard-Jones and embedded-atom method (EAM) potentials, FEM-based continuum mechanics for stress-strain analysis and thermal response evaluation, and ML techniques trained on multiscale descriptors (e.g., bond energy, stress tensor, coordination number) to model nonlinear property relations and accelerate design iteration. We also introduce hierarchical feature fusion modules that enhance learning by combining low-level atomistic descriptors with high-level continuum features, improving both precision and transferability. This approach significantly reduces computational costs and accelerates material design workflows by predicting properties with high precision across a wide range of materials. Our experimental results demonstrate that this integrated methodology outperforms conventional methods in both prediction speed and accuracy, particularly in complex or multicomponent systems, making it a promising tool for future materials research. Benchmark evaluations show improved performance 1 Sample et al.in predicting elastic modulus, thermal conductivity, and phase transition temperature across five material classes. This research aligns with current trends in leveraging advanced digital technologies to enhance materials discovery, offering a robust, scalable, and extensible framework for the optimization and design of advanced materials in various industrial, technological, and scientific applications.