| It’s of vital importance to develop reaction kinetics for good prediction capabilities to describe complex reaction systems and guide pyrolysis technologies. Normally, certain parametric set serves only a certain feeding content in traditional lumping kinetic model. To enhance prediction ability of kinetic model to simulate various feedstocks with less reaction parameters, this thesis probes the influence of molecular structure on reaction processes based on continuous lumping kinetic model at free radical level. Then a molecular structure-oriented continuous lumping kinetic can achieving relatively accurate prediction of their reaction courses.To widen the knowledge of molecular structure influence on reaction processes, a new analytical method and an experimental design are proposed specially for free radical level research. Using published kinetic parameter data of short chain molecules, calculative analyses are performed of molecular structural impact on kinetic constants in free radical reactions. As compensation, cracking reactions of various long chain alkanes are experimentally investigated by Pyrolysis-Gas Chromatography-Mass Spectrum system. As a result, molecular structure significantly influences the reaction processes, through three structural parameters of Chain-length, branching degree and side-chain position, which have effects on pyrolysis courses and their productive profiles by affecting reaction rates and paths.According to research outcomes about molecular structure impact on reaction rates, three parametric functions representing molecular structural influence on reactivity k are introduced into the distribution function D(k), that is, L(k) for that from chain length, F(k) from degree of branching and S(k) from and side-chain position. Yield distribution function p(k,K) is also revised to better reflect characteristics of pyrolytic reactions. The above mentioned improvements and revisions give birth to a novel molecular structure oriented continuous lumping model. It comes out that, the new distribution function, by describing reaction k with the three molecular structural parameters, leads to rational data in accordance with experimental counterparts.Considering a different pattern of the new model, corresponding numerical solution methods are suggested. By familiarizing lumping number features for the three parameters and referencing to current lumped physical property analytical methods, PY-GC/MS system is methodologically borrowed, as online reactor with analytical system, to obtain productive profile. This method then helps to get lumping mass fractions at different moments when take reactivity k as the parameter to make lumps. After gaining reactive kinetic parameters by objective function, the novel numerical solution pathway is ready to the new molecular structure oriented model.Accuracy and applicability certification for different feedstocks of the new model is then carried out. Yields of lumpings with different reactivities are calculated out, in agreement with corresponding experimental ones with errors less than 5%. A 3-parameter olefin productive prediction equation proposal is finished after calculating mass fractions of lumping. The ethylene yield from literature oil is predicted by this method which in accordance with reported exprimental ones. Paramenters from the new continuous lumping equations contribute to an equation to calculate single molecular reactivity. This equation, along with molecular structural parameters, is used in single molecular reactivity computing, with the prediction results in accordance with the experimental ones. These two facts certificate that the new model can reflect reactivity of oil at molecular level and is applicable from different feedstock cases. The new kinetic model can significantly simplify reaction networks and limit parameters, along with applicability for productive profile prediction of several light oils. |
PDF Full Text Request