Bridging across scales from atomistics to the macroscale: 20 years of quasicontinua
Dennis Kochmann, ETH Zürich, Zürich, Switzerland
One of the essential challenges in computational material modeling lies in the wide spectrum of length and time scales to cover, from angstroms and femtoseconds at the atomic level all the up to the macroscopic scales of technological applications. The quasicontinuum (QC) method was introduced as a multiscale technique that bridges across this gap, with the aim of exporting atomistic accuracy to significantly larger scales. Over the 20+ years since its inception, the QC methodology has been extended in numerous directions and has been applied to a myriad of crystalline materials systems. Here, we will review the underlying challenges of atomistic-to-continuum coupling techniques and discuss the QC strategies that enable the bridging across both length and time scales. Unlike brute-force atomistic calculations, this is achieved by a combination of concepts from atomistics, statistical mechanics, continuum mechanics, and computational science. Especially the latter allows for large-scale calculations of superior efficiency. After a survey of the state of the art (with a focus on metallic systems), we present our own flavor within the family of QC methods, which uses intelligent ensemble sampling for minimal force artifacts (which typically arise from a handshake between atomistic and continuum regions) and the principle of maximum entropy for long-term finite-temperature calculations. Finally, we demonstrate how analogous techniques can also be applied to the topical area of structural metamaterials, in which the QC coarse-graining methodology again helps overcome computational length scale constraints.
Session T1: Tuesday, 26 June 2018
End: 09:45 a.m.