Uncertainty Quantification in Multiscale Simulation of Materials: A Prospective

Abstract

Simulation has long since joined experiment and theory as a valuable tool to address materials problems. Analysis of errors and uncertainties in experiment and theory is well developed; such analysis for simulations, particularly for simulations linked across length scales and timescales, is much less advanced. In this prospective, we discuss salient issues concerning uncertainty quantification (UQ) from a variety of fields and review the sparse literature on UQ in materials simulations. As specific examples, we examine the development of atomistic potentials and multiscale simulations of crystal plasticity. We identify needs for conceptual advances, needs for the development of best practices, and needs for specific implementations.