Skip to main navigation Skip to search Skip to main content

Harnessing fluctuations to discover dissipative evolution equations

  • Xiaoguai Li
  • , Nicolas Dirr
  • , Peter Embacher
  • , Johannes Zimmer
  • , Celia Reina
  • University of Pennsylvania
  • Cardiff University
  • University of Warwick

Research output: Contribution to journalArticlepeer-review

17   Link opens in a new tab Citations (SciVal)
104 Downloads (Pure)

Abstract

Continuum modeling of dissipative processes in materials often relies on strong phenomenological assumptions, as their derivation from underlying atomistic/particle models remains a major long-standing challenge. Here we show that the continuum evolution equations of a wide class of dissipative phenomena can be numerically obtained (in a discretized form) from fluctuations via an infinite-dimensional fluctuation-dissipation relation. A salient feature of the method is that these continuum equations can be fully pre-computed, enabling macroscopic simulations of arbitrary admissible initial conditions, without the need of any further microscopic simulations. We test this coarse-graining procedure on a one-dimensional non-linear diffusive process with known analytical solution, and obtain an excellent agreement for the density evolution. This illustrative example serves as a blueprint for a new multiscale paradigm, where full dissipative evolution equations — and not only parameters — can be numerically computed from lower scale data.
Original languageEnglish
Pages (from-to)240-251
Number of pages12
JournalJournal of the Mechanics and Physics of Solids
Volume131
Early online date18 Jun 2019
DOIs
Publication statusPublished - 1 Oct 2019

Keywords

  • Coarse-graining
  • Fluctuation-dissipation
  • Multiscale
  • Non-equilibrium thermodynamics

ASJC Scopus subject areas

  • Condensed Matter Physics
  • Mechanics of Materials
  • Mechanical Engineering

Fingerprint

Dive into the research topics of 'Harnessing fluctuations to discover dissipative evolution equations'. Together they form a unique fingerprint.

Cite this