Cluster Statistical Thermodynamics (CST) — To Efficiently Calculate Quasi-Static Deformation at Finite Temperature Based on Molecular Potential

Abstract

For quasi-static deformations in engineering practice, molecular dynamics (MD) simulation requires computation resources that are not affordable even with ever-increasing computing power. In order to overcome this weakness, we developed a new method called cluster statistical thermodynamics (CST). By taking the advantage of statistical thermodynamics and adopting finite-element interpolation, the new approach can not only simulate quasi-static deformation but have very high computing efficiency. The new method is based on molecular potentials as MD does, but statistical thermodynamics help us greatly reduce the tedious calculation of thermal fluctuations of molecules. Therefore, the new method appears to be superior to MD in the simulations of quasi-static deformation. Especially CST works much more efficiently than MD with much less storage space and CPU time. In this paper, we illustrate the new methodology by means of some examples of two-dimensional quasi-static tensile process at 300 K. It is found that the results obtained with CST are in good agreement with those obtained by fully atomistic simulations and CST is 600 times faster than MD. Hence, the new method seems to be a very efficient and promising approach to numerical simulations of solid deformations under quasi-static loadings and at finite temperatures, based on molecular potentials.