Numerical Investigation into the Formation of CO and Oxygenated and Nonoxygenated Hydrocarbon Emissions from Isooctane- and Ethanol-Fueled HCCI Engines

Abstract

The present study focuses on the investigation of the formation of CO and unburned oxygenated and nonoxygenated hydrocarbon emissions from HCCI engines fueled with neat ethanol and neat isooctane. This is achieved with the use of a multizone model, which describes the essential features of HCCI combustion, that is, heat and mass transfer within the combustion chamber, both of which are modeled using phenomenological submodels. These mechanisms affect the formation of the main HCCI engine pollutants, namely, unburned hydrocarbons and carbon monoxide. Combustion is simulated using chemical kinetics coupled to oxidation mechanisms for isooctane and ethanol. These mechanisms also describe the decomposition of the original fuel into intermediate hydrocarbons and carbon monoxide. A validation of the model for both fuels is given for various load cases. In the numerical investigation, the formation of CO is described for the corresponding experimental cases and the essential features of the transition from CO production due to bulk quenching and to CO production due to postcombustion partial HC oxidation are shown. Additionally, the formation of HC emissions is described including both oxygenated and nonoxygenated compounds. This distinction was found to be necessary since both fuels include oxygenated species in the exhaust gases, the relative amount of which depends on load conditions and the fuel used. The fraction of oxygenated compounds to total unburned HC is high for ethanol at all loads, primarily due to the presence of ethanol, acetaldehyde and formaldehyde, in descending order of importance. The relative proportion of oxygenates in total unburned HC in the case of isooctane was found to depend on load. These findings raised questions regarding the assessment of unburned hydrocarbon emissions using conventional measuring devices, such as the FID. For this reason the relative error in the FID measurement was estimated, using the simulated HC composition results and the FID relative response of each of the species constituting the HC.