Integration of On-The-Fly Kinetic Reduction with Multidimensional CFD Kaiyuan He, Marianthi G. Ierapetritou, and Ioannis P. Androulakis Dept of Chemical and Biochemical Engineering, Rutgers, The State University of New Jersey, Piscataway, NJ 08854 DOI 10.1002/aic.12072 Published online September 29, 2009 in Wiley InterScience (www.interscience.wiley.com). A reduction approach for coupling complex kinetics with engine computational fluid dynamics (CFD) code has been developed. An on-the-fly reduction scheme was used to reduce the reaction mechanism dynamically during the reactive flow calculation in order to couple comprehensive chemistry with flow simulations in each computational cell. KIVA-3V code is used as the CFD framework and CHEMKIN is employed to for- mulate chemistry, hydrodynamics and transport. Mechanism reduction was achieved by applying element flux analysis on-the-fly in the context of the multidimensional CFD calculation. The results show that incorporating the on-the-fly reduction approach in CFD code enables the simulation of ignition and combustion process accurately compared with detailed simulations. Both species and time-dependant information can be provided by the current model with significantly reduced CPU time. V V C 2009 American Institute of Chemical Engineers AIChE J, 56: 1305–1314, 2010 Keywords: on-the-fly reduction, CFD, detailed kinetics Introduction As the increase of computational capability in recent years, more accurate computational fluid dynamics (CFD) models have been developed. The fundamental basis of any CFD model is the solution of Navier-Stokes equations. 1 Further complicating factors are that the flow occurs in a complex geometry, the incorporation of turbulence, and the integration of detailed chemical kinetics. To describe turbu- lence several models have been developed such as direct nu- merical simulation (DNS) method 2 and k-e model. 3 In recent years much progress has been made in CFD model develop- ment for engines. For example, studies have been conducted for direct-injection diesel engines, 4,5 indirect-injection diesel engines, 6 stratified-charge rotary engines, 7 and homogene- ous-charge compression ignition engines. 8,9 Although these engine simulation codes are comprehensive and can predict engine details to some extent, they are not entirely predictive for the combustion process due to the wide range of combus- tion time scales and engine conditions. Thus, submodels have been developed to capture the short time scale proc- esses such as drop vaporization 10–12 and turbulence disper- sion. 13,14 A detailed review of CFD packages and their capa- bility is given in. 15 CFD code packages have been developed to include both the fluid dynamics models and these complementary submo- dels for engine simulation. However, most of these packages are for commercial use and the source codes are generally not available. In this study, KIVA code 16 has been selected as the engine model, since it is one of the most commonly used CFD packages and has the ability to calculate three- dimensional (3-D) flows in engine cylinders, arbitrary piston shapes, and capture the effects of turbulence and heat gra- dients near walls. The package has been widely used in engine simulations to predict ignition and combustion pro- cess. 17–19 The more essential factor is that the source code of KIVA is accessible and thus it can be used in model de- velopment work as a test bed. 17 Given the model’s capability to capture detailed flow properties, it has also been widely recognized that incorporating detailed chemistry in CFD cal- culation to describe a reactive flow is necessary. The realiza- tion of this has motivated continuing development of CFD Correspondence concerning this article should be addressed to I. P. Androulakis at yannis@rci.rutgers.edu V V C 2009 American Institute of Chemical Engineers AIChE Journal 1305 May 2010 Vol. 56, No. 5