A continued challenge to engine combustion simulation is predicting the impact of fuel-composition variability on performance and emissions. Diesel fuel properties, such as cetane number, aromatic content and volatility, significantly impact combustion phasing and emissions. Capturing such fuel property effects is critical to predictive engine combustion modeling. In this work, we focus on accurately modeling diesel fuel effects on combustion and emissions. Engine modeling is performed with 3D CFD using multi-component fuel models, and detailed chemical kinetics. Diesel FACE fuels (Fuels for Advanced Combustion Engines) have been considered in this study as representative of street fuel variability. The CFD modeling simulates experiments performed at Oak Ridge National Laboratory (ORNL)  using the diesel FACE fuels in a light-duty single-cylinder direct-injection engine. These ORNL experiments evaluated fuel effects on combustion phasing and emissions. The actual FACE fuels are used directly in engine experiments while surrogate-fuel blends that are tailored to represent the FACE fuels are used in the modeling. The 3D CFD simulations include spray dynamics and turbulent mixing.
We first establish a methodology to define a model fuel that captures diesel fuel property effects. Such a model should be practically useful in terms of acceptable computational turnaround time in engine CFD simulations, even as we use sophisticated fuel surrogates and detailed chemistry. Towards these goals, multi-component fuel surrogates have been developed for several FACE fuels, where the associated kinetics mechanisms are available in a model-fuels database. A surrogate blending technique has been employed to generate the multi-component surrogates, so that they match selected FACE fuel properties such as cetane number, chemical classes such as aromatics content, T50 and T90 distillation points, lower heating value and H/C molar ratio. Starting from a well validated comprehensive gas-phase chemistry, an automated method has been used for extracting a reduced chemistry that satisfies desired accuracy and is reasonable for use in CFD. Results show the level of modeling necessary to capture fuel-property trends under these widely varying engine conditions.