On molecular modelling and continuum concepts (Q5962113)

The paper examines the derivation and physical interpretation of candidate interaction-stress tensors and draws attention to the manner in which continuum fields and concepts can be linked with scale-independent molecular averages. First, linear balance is derived on the base of an equation governing the motion of a molecule in an inertial frame, taking into account the forces exerted on the molecule by other molecules. After introducing general requirements for a weighting function \(w\) and modeling molecules as point masses, mass and momentum densities are determined taking into account that the systems are considered whose molecular content does not change with time. This is carried out for any choice of \(w\) and shows a good satisfaction of the continuity equation. Several local forms of linear momentum balance are defined, each involving a kinetic stress. For candidate interaction stresses, the author derives an equation including an interaction force density, and the corresponding Cauchy stresses are derived as differences between the candidate interaction stresses and the kinetic stress. For each choice of \(w\), the region occupied by a body, its motion and material points are derived by scale-dependent constructions. Then a number of various candidate tensors is discussed. Specific choices of \(w\) are considered (all of them give a rise to scale-dependent fields), several of which have particular physical merit. Physical interpretations are given for the simplest choice of \(w\) for which any velocity field is identified with a local mass center motion. It is shown that the kinetic stress a has thermal character, and explicit forms of candidate interaction stresses are demonstrated and compared. Temporal averaging is implemented to obtain standard local forms of the continuity equation and of the linear momentum balance. All terms therein are related to space-time averages computed at any given scales of length and time. Then, another approach is presented: microscopic regions and their boundaries are partitioned into specific cells, and sub-surfaces and balances are obtained in the global form constituting the starting point for standard continuum modeling. Terms in these balances are related to scale-dependent cellular averages. Finally, the author makes some remarks on the moment of momentum and energy balances, on the comparison of weighting function and cellular averaging, on future studies of behavior at boundaries/interfaces, and on the molecular interpretation of configurational forces.