Perspectives on the design and control of multiscale systems

Carbon nanotube growth: First-principles-based kinetic Monte Carlo model

“…Previous experimental and theoretical studies have reported that the activation energies associated with Ni catalytic methane cracking are in the range of 0.91-1.01 eV [98-100] and 1.06-1.46 eV [101][102][103][104][105][106], respectively. These discrepancies in the results illustrate the uncertainties in the DFT estimation of activation energies which translates into variability in the predicted CNT growth rate; this is a key emerging subject in multi-scale modeling [107][108][109][110]. To examine the sensitivity of the predicted CNT growth rates on the energetics used in the AKMC simulations, the uncertainty associated with the energetics in the critical elementary steps was studied in this work.…”

Simplified kinetic modelling and numerical simulation of a metal oxide chemical bath electro deposition process at a rotating electrode

“…To estimate the local pH increasing or the deposition velocity the privileged way is the numerical calculation [10][11][12][13][14]. Many works have been performed to model and to numerically simulate electro deposition process, from macroscopic scale to mesoscopic scale.…”

A robust nonlinear model predictive controller for a multiscale thin film deposition process

“…It is, however, not trivial to evaluate accurate model parameters at molecular scale (Evans and Ricardez-Sandoval, 2014). Although multiscale processes mostly encompass fine-scale features that cannot be known with absolute certainty, model-plant mismatch has mostly been ignored in the proposed methodologies in the literature (Braatz et al, 2006). For uncertainty analysis of continuum models, analytical techniques have been proposed to characterize the distribution or worst-case deviation of the control performance under model parameter uncertainty (Nagy and Braatz, 2003a).…”

“…In other approaches, the lack of knowledge at the fine-scale has been addressed by optimal experimental designs for parameter optimization (Prasad and Vlachos, 2008). To avoid high computational simulations of the finescale model, these models are approximated using PSEs (Braatz et al, 2006;Raimondeau et al, 2003).…”