A method and system for predicting thermodynamic properties of hexafluoride uranium in the whole phase region

Abstract

This application provides a method and system for predicting the thermodynamic properties of uranium hexafluoride in its entire phase region, relating to the field of nuclear chemical process simulation technology. The method first acquires the all-electron potential energy surface data of the target molecule through full-space potential energy surface scanning, and then uses a Boltzmann weighting factor to clean the all-electron potential energy surface data with high-energy noise to generate macroscopic force field parameters. Then, based on a gradient-aware CPU and GPU collaborative architecture, computational content is allocated, and the gas-liquid equilibrium evolution of liquid film construction and vacuum-induced evaporation processes is executed to obtain thermodynamic property data. This method can predict the thermodynamic properties of high-hardness, strongly sterically hindered molecules such as uranium hexafluoride in its entire phase region with resistance to numerical collapse. By establishing a cross-scale coupling mechanism across micro, meso, and macro scales, it achieves long-term stable and high-precision prediction of the properties of the target molecule in its entire phase region, from low-temperature solid to high-temperature supercritical fluid.

Images