ITR/AP (CHE): Mutliscale Treatment of Fluid-Solid Interfaces: Development of Hybrid Monte Carlo and Finite Element Code
University Of Nebraska-Lincoln, Lincoln NE
Investigators
Abstract
0112929 Zheng This ITR/AP Small Group Award supports development of efficient and robust computational methods for simulating physical systems involving phenomena at drastically different length scales. The results will serve an important need in computer-aided design, analysis and modeling. A classic example is a system that involves fluid-solid interfaces with complex geometries on scales ranging from molecular to macroscopic. On one hand, computational modeling approaches based on continuum approximations (e.g., finite-element and finite -difference techniques) are capable of describing mechanical responses only at wavelengths much longer than the typical distances over which the fluid density at interfaces varies appreciably. On the other hand, the whole system is far too large to be treated by a fully atomistic simulation that would encompass responses at all scales simultaneously. To address this problem a hybrid Monte Carlo-finite element scheme, in which fluid molecules at the interface are treated explicitly, while the solid is handled by continuum techniques, will be applied. A "reduced" description in which the stochastic variables consist of the positions of the fluid molecules and the nodes of the finite-element mesh covering the solid phase will be developed. The stochastic nature of Monte Carlo lends itself to an efficient parallel computing scheme, which is crucial for simulating real-world systems. A multi-disciplinary team has been assembled to undertake the three-year project, the expected outcome of which is a powerful computational scheme for analyzing the thermomechanical response of realistic fluid-solid systems in which the coupling of the molecularly heterogeneous interfaces to the macroscopic continuum plays an essential role. ***
View original record on NSF Award Search →