In order to obtain a deeper understanding of the performance of a diesel oxidation catalyst (DOC), techniques are developed which help to bridge the gap in the development of catalyst technology from the laboratory bench-top to an application on a real 2.0 litre diesel engine. The methodology is illustrated using a full-scale DOC (106 mm diameter; length 114 mm), from which sections are cut and then examined using a variety of techniques, including: optical measurement, SEM, TEM, FTIR, CO chemisorption, pore size and pore volume measurements. These provide valuable information on the characteristics of the Pt catalyst, in the actual form (on supportlwashcoat) that it will be used in a commercial application. In addition, a transient chromatographic technique was tested to measure the effective diffusivity in the washcoat layer. These experiments were performed using sections of v-alumina washcoated (:::::50/lm thick) monolith (1 mm channels), which were mounted in a 570 mm long housing - unfortunately this did not work. The technique was improved, and short sections of monolith, with a larger channel diameter (2.64 mm) and with a thicker washcoat (:::::475/lm) were mounted in a 1 m long sleeve - although better results were obtained, this technique still needs to be improved. A methodology was then described, in which a short section of DOC (5 mm long) was cut from the full-scale DOC, and this was connected to a live 2.0 litre diesel engine. It was shown how experiments (with gas inlet temperatures varying from 150 to 270 QC) can be performed, and in combination with a modelling technique, the information gathered can be used to test the validity of rate expressions. It was shown: (a) that the Langmuir-Hinshelwood-Hougen-Watson (LHHW) form of rate expression is clearly unsuitable for transient simulations, whereas mechanistic forms can provide a better match to the experimental data, and (b) how to tune the coefficients to match points of ignition-extinction in the light-off hysteresis cycle.

This research was aimed at providing a better understanding of regeneration processes in wall-flow diesel particulate filters (DPFs), with emphasis on the combustion of particulate matter (PM). A 1-D model was used to investigate the effects of inherent PM properties on DPF regeneration behaviour. These properties were mean particulate diameter, porosity and bulk density of the PM, as well as the kinetic parameters of PM oxidation, i.e. frequency factor and activation energy. A parametric study showed that the activation energy of the PM oxidation reaction was the most important parameter and this was followed by the associated frequency factor, bulk density and porosity and mean particulate diameter. Due to the importance of the kinetic parameters of the PM oxidation reactions, a new 1-D model with a multi-step reaction scheme that required no tuneable kinetic parameters for the PM oxidation reactions was developed.

Hydrotreating processes play a vital role in petroleum refineries to meet the increasing demand of transportation fuels. The recent trends in processing of heavier crudes with higher sulphur contents and more stringent product specifications for cleaner transportation fuels, such as ultra-low sulphur diesel, are resulting in more severe operating conditions and higher hydrogen consumption of hydrotreating processes. In order to carry out any revamp or design projects for improving the performance and efficiency of hydrotreating units molecular kinetic models of hydrotreating reactions may be required to provide detailed and accurate information of the composition and properties of hydrotreating products. The overall hydrotreating process consisting of the hydrotreater, the separation system and the associated heat recovery system need to be modelled on a consistent basis of detailed characterisation of petroleum fractions and the interactions of these individual subsystems with each other and with the refinery hydrogen network handled simultaneously for overall process optimisation. A molecular pathways level model of diesel hydrotreating reactions using the molecular type and homologous series matrix is employed for prediction of detailed molecular level information of composition and properties of diesel hydrotreating products. The molecular type and homologous series matrix representation of petroleum fractions is a detailed characterisation approach that represents the composition of a stream in a matrix in terms of the carbon number range and compound classes existing in the petroleum fraction. A new strategy is developed for estimation of physical properties of middle distillate and heavy petroleum fractions with molecular type and homologous series matrix representation using group contribution methods.

This work is concerned with the development of a new quantitative kinetic model for the analysis of hydrocarbon fuel droplet heating and evaporation, suitable for practical engineering applications. The work mainly focuses on the following two areas. Firstly, a new molecular dynamics (MD) algorithm for the simulation of complex hydrocarbon molecules, with emphasis on the evaporation/condensation process of liquid n-dodecane (C12H26), which is used as an approximation for Diesel fuel, has been developed. The analysis of n-dodecane molecules has been reduced to the analysis of simplified molecules, consisting of pseudoatoms, each representing the methyl (CH3) or methylene (CH2) groups. This analysis allows us to understand the underlying physics of the evaporation/condensation process of n-dodecane molecules and to estimate the values of its evaporation/condensation coefficients for a wide range of temperatures related to Diesel engines. Nobody, to the best of our knowledge, has considered MD simulation of molecules at this level of complexity. Secondly, a new numerical algorithm for the solution of the Boltzmann equation, taking into account inelastic collisions between complex molecules, has been developed. In this algorithm, additional dimensions referring to inelastic collisions have been taken into account alongside three dimensions describing the translational motion of molecules as a whole. The conservation of the total energy before and after collisions has been considered. A discrete number of combinations of the values of energy corresponding to translational and internal motions of molecules after collisions have been allowed and the probabilities of the realisation of these combinations have been assumed to be equal. This kinetic model, with the values of the evaporation coefficient estimated based on MD simulations, has been applied to the modelling of the heating and evaporation processes of n-dodecane droplets in Diesel engine-like conditions. In the previously developed kinetic models, applied to this modelling, all collisions were assumed to be elastic and the evaporation coefficient was assumed equal to 1. It is shown that the effects of inelastic collisions lead to stronger increase in the predicted droplet evaporation time relative to the hydrodynamic model, compared with the similar increase predicted by the kinetic model considering only elastic collisions. The effects of a non-unity evaporation coefficient are shown to be weak at gas temperatures around or less than 1,000 K but noticeable for gas temperatures 1,500 K. The application of the rigorous kinetic model, taking into account the effects of inelastic collisions and a non-unity evaporation coefficient, and the model considering the temperature gradient inside droplets is recommended when accurate predictions of the values of droplet surface temperature and evaporation time in Diesel engine-like conditions are essential.