Adaptive Carbon Monoxide Kinetics for Exhaust Aftertreatment Modeling

Future emission standards are driving the need for advanced control of both Spark (SI) and Compression Ignition (CI) engines. However, even with the implementation of cooled Exhaust Gas Recirculation and Low Temperature Combustion (LTC), it is unlikely that in-cylinder combustion strategies alone will reduce emissions to levels below the proposed standards. As a result, researchers are developing complex catalytic aftertreatment systems to meet these tailpipe regulations for both conventional and alternative combustion regimes. Simulating these exhaust systems requires fast and accurate models suitable for significant changes in inlet conditions. Most aftertreatment devices contain Platinum Group Metals because of their widely documented beneficial catalysis properties; examples include Diesel Oxidation Catalysts, Three-Way Catalysts and Lean NO_x Traps. There are kinetic mechanisms available for each of these devices, but often they do not extrapolate well to other formulations. For example, Carbon Monoxide (CO) levels entering a catalyst are significantly different between an SI and CI engine. In addition, modifying engine control to utilize LTC operation can result in an increase in CO levels due to lower combustion efficiency. This adversely affects the conversion capabilities of a catalytic device through increased levels of CO inhibition. Finally, catalyst loading and metal dispersion differences between devices often prohibit a direct extension of kinetic constants. As a result, mechanisms often need recalibration for correct modeling capabilities. In order to begin creating a more predictive kinetic mechanism, this paper simulates CO oxidation as a function of different inlet concentration levels and metal loadings. While aftertreatment devices contain many reactions, modeling of one fundamental reaction is a first step to determine the feasibility of adaptive kinetics. In addition, research into the history of the CO oxidation mechanism over platinum illustrates a more accurate rate expression to utilize in deference to current modeling activities. The authors calibrate this expression to experimental data taking into account significant changes in inlet conditions, metal loading and dispersion values. Model fidelity is determined through the simulation of additional data not part of the initial calibration efforts. In addition, the paper discusses strengths and weaknesses of the model along with how other researchers can help foster adaptive kinetic development.