Computer Modeling of HCCI Combustion

Doctoral thesis, 2004

During the last decade an alternative to conventional internal combustion engines, i.e. spark-ignited and Diesel engines, has been under investigation by an ever-increasing number of research groups. This alternative process is called Homogeneous Charge Compression Ignition (HCCI) and engine tests have shown that it has great potential to significantly reduce harmful emissions, and it is expected to reduce fuel consumption by 10-20% compared with spark-ignited operation. HCCI is defined as the process in which a (more or less) homogeneous mixture of air and fuel, diluted with excess air as well as combustion products, is compressed under such conditions that auto-ignition occurs near the end of the compression stroke, followed by a combustion process that is significantly faster than conventional Diesel or Otto combustion. This thesis is concerned with the numerical investigation of this novel process. The principal aim of the work underlying the thesis has been to develop simulation tools that can contribute to the understanding and further development of the HCCI process. Since HCCI combustion is strongly dependent on chemical kinetics, some of the work presented involved the construction of a reaction mechanism that can describe the oxidation of gasoline-like fuels. Because real gasoline is a complex mixture of many components, the model fuel is assumed to be a blend of iso-octane, n-heptane and toluene. The reaction mechanism consists of approximately 100 species and 500 reactions. The construction and validation processes of the reaction mechanism are described and the implementation of the mechanism into different HCCI combustion models is discussed. Several models, with varying levels of complexity and corresponding calculation times, have been developed, validated and used in this analysis. The simplest model presented applies simple expressions to predict auto-ignition timing and heat release during combustion. Due to its short calculation times it is a suitable tool for optimizing combustion control strategies. An example of such an application is presented. More complex simulation tools discussed are the so-called single-zone and multi-zone HCCI models. The developed reaction mechanism is implemented in these models and a detailed-chemistry approach is used to predict the auto-ignition timing and the rate of combustion. Whereas the zero-dimensional single-zone model is based on the assumption of a perfectly homogeneous mixture, the multi-zone model is more sophisticated since multiple zones are used to represent the mixture inhomogeneity present in the combustion chamber of real engines.The combustion models described above were connected to a cycle simulation code, which can be applied to simulations of the complete HCCI engine cycle, i.e. including the gas exchange processes which are important for HCCI mixture preparation. Predictions by the different models are compared with experimental data obtained from HCCI test engines at Chalmers University of Technology (CTH). It is shown that the match between measured and calculated results/trends is generally good. Relative errors for the predictions of load and fuel consumption are around 10%. In summary, the thesis describes the development, validation and application of a range of HCCI simulation tools based on a detailed reaction mechanism for gasoline-like fuels. The model predictions are shown to correspond well to experimental data in general.