Applications of lumping kinetics methodology to complex reactive mixtures

Abstract

The thesis concerns a comprehensive study to develop and assess the predictive capabilities of a lumping model for the kinetics of complex reaction mixtures containing a large number of reacting components. Two types of lumping models were developed to reduce the number of reaction kinetics, namely a discrete and a continuum lumping model. These models were applied on three different problems. In the first case study, a continuum model was developed for a mixture of n-paraffins (waxes) produced from the Fischer-Tropsch process; the paraffins range from C5 to C70 and undergoing catalytic hydrocracking. The model was run with two types of the reactant-type distribution functions to describe the yield of products from the isomerisation and cracking reactions. The model was used to study the effect of the operating conditions on the model parameters and the yield composition. Experimental data were used to optimise the model parameters. The optimal parameters were used to predict the product distribution of n-paraffins hydrocracking and their conversion. The new in this case study was used the carbon number as label in the continuum lumping model and to study how the effect of D(k) on the yield distribution. Good agreements have been obtained when running the model with D(k) employed a gamma function but it needs more time to solve the model than when employing a power law relation for D(k). In the second case study, both primary and secondary reaction kinetics of the pyrolysis of lignin were investigated by using the discrete lumping methodology. Two mathematical models were developed which consider the product as three lumps whilst the lignin was assumed to be an additional lump. The model’s results were validated against experimental data. In addition, a continuum lumping model was developed for the cracking of the tar to obtain lighter components. The novelty in this case study is to develop a kinetic model including primary and secondary reaction kinetics for the pyrolysis of lignin in a fluid bed pyrolyser and to study how the continuum lumping model for tar can be linked to the discrete lumping model. In the third case study, a model based on the continuum lumping approach was proposed to predict the molecular weight distribution of polymers during batch polymerisation. The result obtained from a continuum model was assessed, at this stage, only qualitatively; nevertheless, by analysing the weight distribution and the average of such distribution, conclusions were reached to assess the predictive capability of the lumping methodology. It is the first time that the continuum model with a yield distribution function is used to predict the molecular weight distribution of the polymerisation at various times.