Book Description

Developing a profound understanding of the combustion characteristics of the cold-start phase of a Direct Injection Spark Ignition (DISI) engine is critical to meeting the increasingly stringent emissions regulations. Computational Fluid Dynamics (CFD) modeling of gasoline DISI combustion under normal operating conditions has been discussed in detail using both the detailed chemistry approach and flamelet models (e.g., the G-Equation). However, there has been little discussion regarding the capability of the existing models to capture DISI combustion under cold-start conditions. Accurate predictions of cold-start behavior involves the efficient use of multiple models - spray modeling to capture the split injection strategies, models to capture the wall-film interactions, ignition modeling to capture the effects of retarded spark timings, combustion modeling to accurately capture the flame front propagation, and turbulence modeling to capture the effects of decaying turbulent kinetic energy. The retarded spark timing helps to generate high heat flux in the exhaust for a rapid catalyst light-off of the after-treatment system during cold-start. However, the adverse effect is a reduced turbulent flame speed due to decaying turbulent kinetic energy. Accordingly, developing an understanding of the turbulence-chemistry interactions is imperative for accurate modeling of combustion under cold-start conditions.This study introduces a modified version of the G-Equation combustion model called the GLR model (G-Equation for Lower Reynolds number regimes) that exhibits improved performance under cold-start conditions. The model attempts to estimate the turbulent flame speed based on the local conditions of fuel concentration and turbulence intensity. The local conditions and the associated turbulent-chemistry interactions are studied by tracking the flame front on the Borghi-Peters regime diagram. To accurately model the DISI combustion process, it is important to account for the effects of the spark energy discharge process. In this work, an ignition model is presented that is compatible with the G-Equation combustion model, and which accounts for the effects of plasma expansion and local mixture properties such as turbulence and the equivalence ratio on the early flame kernel growth. The model is referred to as the Plasma Velocity on G-Surface (PVG) model, and it uses the G-surface to capture the kernel growth. The model derives its theory from the DPIK model and applies its concepts onto an Eulerian framework, thereby removing the need for Lagrangian particles to track the kernel growth. Finally, a methodology of using machine learning (ML) techniques in combination with 3D CFD modeling to optimize the cold-start fast-idle phase of a DISI engine is presented. The optimization process implies the identification of the range of operating parameters, that will ensure the following criteria under cold-start conditions: (1) a fixed IMEP of 2 bar (BMEP of 0 bar), (2) a stoichiometric exhaust equivalence ratio (based on carbon-to-oxygen atoms) to ensure the efficient operation of the after-treatment system, (3) enough exhaust heat flux to ensure a rapid light-off of the after-treatment system, and (4) acceptable NOx and HC emissions. Gaussian Process Regression (GPR)-based ML models are employed to make predictions about DISI cold-start behavior with acceptable accuracy and a substantially reduced computational time.