We propose a data-driven framework for rapid prediction of thermal conductivity in solids based on shorttime molecular dynamics(MD)simulations.By converting atomic configurations into graph representations,a graph con...We propose a data-driven framework for rapid prediction of thermal conductivity in solids based on shorttime molecular dynamics(MD)simulations.By converting atomic configurations into graph representations,a graph convolutional network(GCN)is used to extract spatial features,which are then processed by a long short-term memory(LSTM)network to capture the temporal evolution of physical properties.The framework is validated using equilibrium MD simulations of germanium at 1000 K across various system sizes.With sizespecific normalization and optimized hyperparameters,the model accurately predicts the converged thermal conductivity,achieving results consistent with experimental data.Notably,the proposed method significantly reduces computational time by up to 800-fold at large system sizes,which demonstrates its potential to accelerate thermal transport simulations in solid-state systems.展开更多
基金supported by the National Natural Science Foundation of China(Grant No.52376063)the High Performance Computing Center,Yangtze Delta Region Academy in Jiaxing,Beijing Institute of Technology。
文摘We propose a data-driven framework for rapid prediction of thermal conductivity in solids based on shorttime molecular dynamics(MD)simulations.By converting atomic configurations into graph representations,a graph convolutional network(GCN)is used to extract spatial features,which are then processed by a long short-term memory(LSTM)network to capture the temporal evolution of physical properties.The framework is validated using equilibrium MD simulations of germanium at 1000 K across various system sizes.With sizespecific normalization and optimized hyperparameters,the model accurately predicts the converged thermal conductivity,achieving results consistent with experimental data.Notably,the proposed method significantly reduces computational time by up to 800-fold at large system sizes,which demonstrates its potential to accelerate thermal transport simulations in solid-state systems.
