Applying uncertainty quantification to multiphase flow computational fluid dynamics

A. Gel, R. Garg, C. Tong, M. Shahnam, C. Guenther

Research output: Contribution to journalArticle

29 Scopus citations

Abstract

Multiphase computational fluid dynamics plays a major role in design and optimization of fossil fuel based reactors. There is a growing interest in accounting for the influence of uncertainties associated with physical systems to increase the reliability of computational simulation based engineering analysis. The U.S. Department of Energy's National Energy Technology Laboratory (NETL) has recently undertaken an initiative to characterize uncertainties associated with computer simulation of reacting multiphase flows encountered in energy producing systems such as a coal gasifier. The current work presents the preliminary results in applying non-intrusive parametric uncertainty quantification and propagation techniques with NETL's open-source multiphase computational fluid dynamics software MFIX. For this purpose an open-source uncertainty quantification toolkit, PSUADE developed at the Lawrence Livermore National Laboratory (LLNL) has been interfaced with MFIX software. In this study, the sources of uncertainty associated with numerical approximation and model form have been neglected, and only the model input parametric uncertainty with forward propagation has been investigated by constructing a surrogate model based on data-fitted response surface for a multiphase flow demonstration problem. Monte Carlo simulation was employed for forward propagation of the aleatory type input uncertainties. Several insights gained based on the outcome of these simulations are presented such as how inadequate characterization of uncertainties can affect the reliability of the prediction results. Also a global sensitivity study using Sobol' indices was performed to better understand the contribution of input parameters to the variability observed in response variable.

Original languageEnglish (US)
Pages (from-to)27-39
Number of pages13
JournalPowder Technology
Volume242
DOIs
StatePublished - Jul 1 2013

Keywords

  • Computational fluid dynamics (CFD)
  • Data-fitted response surface
  • Multiphase flow
  • Non-intrusive parametric uncertainty quantification and propagation
  • Surrogate models

ASJC Scopus subject areas

  • Chemical Engineering(all)

Fingerprint Dive into the research topics of 'Applying uncertainty quantification to multiphase flow computational fluid dynamics'. Together they form a unique fingerprint.

  • Cite this