Evaluation of the Predictive Capabilities of a Phenomenological Combustion Model for Natural Gas SI Engine

Abstract

The current study evaluates the predictive capabilities of a new phenomenological combustion model, available as a part of the GT-Suite software package. It is comprised of two main sub-models: 0D model of in-cylinder flow and turbulence, and turbulent SI combustion model. The 0D in-cylinder flow model (EngCylFlow) uses a combined K-k-ε kinetic energy cascade approach to predict the evolution of the in-cylinder charge motion and turbulence, where K and k are the mean and turbulent kinetic energies, and ε is the turbulent dissipation rate. The subsequent turbulent combustion model (EngCylCombSITurb) gives the in-cylinder burn rate; based on the calculation of flame speeds and flame kernel development. This phenomenological approach reduces significantly the overall computational effort compared to the 3D-CFD, thus allowing the computation of full engine operating map and the vehicle driving cycles. Model was calibrated using a full map measurement from a turbocharged natural gas SI engine, with swirl intake ports. Sensitivity studies on different calibration methods, and laminar flame speed sub-models were conducted. Validation process for both the calibration and sensitivity studies was concerning the in-cylinder pressure traces and burn rates for several engine operation points achieving good overall results.

Keywords

• Predictive Phenomenological Model
• Internal Combustion Engine
• Spark-Ignition
• K-k-ε Kinetic Energy Cascade
• 0D In-Cylinder Flow Model
• Turbulent SI Combustion Model
• Natural Gas Engine
• Genetic Algorithm
• Gt-Suite