MRI: Acquisition of a High Performance Beowulf Cluster Computer for Materials Modeling

Lehigh University, Bethlehem PA

Investigators

Abstract

This proposal requests funds for a medium- sized parallel cluster computer to investigate various computationally intensive problems arising in materials science and mechanics. A particular problem of interest is the computation of the mesoscale dynamics of systems of dislocations, and their relation to continuum plastic deformation. Another problem that is the focus of current work is the estimation of local elastic properties using mean square displacement data from a large parallel molecular dynamics code. Other ongoing efforts that will benefit from this facility are the simulation of solute diffusion to a crack tip using kinetic Monte Carlo techniques and the effect of dopants upon the creep resistance of high strength ceramics. Computational materials modeling at Lehigh University is an interdisciplinary enterprise involving researchers in several allied departments. Much of the ongoing work is focused on obtaining a better understanding of how various structural features affect the mechanical properties of materials over a wide range of length and time scales. Indeed, research projects ongoing at Lehigh have begun to produce significant advances, for example, in our understanding of the high-temperature creep behavior of ceramic oxides, in the development of constitutive relations for plasticity that link the mesoscale and the continuum, in the quantification of local stresses and elastic constants, and in the description of fracture behavior in materials with segregating impurities. This proposal requests funds to acquire a relatively large Beowulf computing cluster that will enable us to make significant leaps in our research program in computational materials modeling