CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Main CFD Forum (
-   -   Finite Element vs Molecular Dynamics (

John August 22, 2002 22:01

Finite Element vs Molecular Dynamics
How accurate is if we simulate ' a single array of molecules on a bulk surface' with finite element analysis implemented codes?

Where do finite element analysis and molecular dynamic simulation meet?

Is there any hybrid codes available?



marco August 29, 2002 11:36

Re: Finite Element vs Molecular Dynamics
Finite elemnt is a macro method solving continuum problems. MD is a purely microscopic simulation of LJ atoms or molecules. You question deals with the coarse-graining problem in physics, which means how can we reduce the freedom degree of the physical problems retaining all the relavant physical informations. Many mesoscopic techniques have been developed in the past ten years for example DPD (Dissipative Particle Dynamics). This is a "molecular dynamics-like" simulation, where every particle represents, let's say, a cluster of 100 LJ molecules, and the interaction between them are modelled with a conservative+disspative+stochastic force. In the case of fluids, for example, the correct Navier-Stokes solution are obtained by Chapman-Enskog expansion. On bigger scale DPD meet SPH (Smoothed Particle Hydrodynamics) which is still a "MD-like" numerical scheme but it deal with bigger scale (it is on the same level of FEM but but it has the structure of a Lagrangian particle method). I'm trying to study this problem for metals. "How can I set up few parameters of a SPH viscoelatic code, working on microscale, in order to set up resulat from MD onb nanoscale". The comparison is still under work, I'm studyng simple shear-flows, with or without fixed obstacles and comparing the viscometric functions with thoose coming from MD. If you are interesting in some references or literature tell me.

All times are GMT -4. The time now is 17:38.