Analysis of the Impact of Uncertainties in Inputs on CFD Predictions of Gasoline Compression Ignition

Abstract

Computational fluid dynamics (CFD) is a valuable tool to gain insights into the combustion process, particularly for novel engine combustion concepts that do not have significant experimental data available. However, prediction of targets of interest from a CFD simulation can oftentimes be quite sensitive to the uncertainties in inputs to the CFD model. These uncertainties could be in the experimental boundary and initial conditions, fuel properties, CFD model constants, chemical kinetic rates, etc. In this work we isolate the effect of uncertainties in some key inputs in the form of experimental boundary conditions and CFD model parameters on combustion and emissions targets of interest for gasoline compression ignition (GCI) at two operating conditions — idle, and low-load. The uncertainties in the subset of inputs studied in this work were identified to have the greatest impact out of 34 inputs to the CFD model studied by means of a global sensitivity analysis (GSA) performed in our prior work. The goal of this study is to perform a more focused study as a follow-on to that GSA, by perturbing only a single parameter at a time.