Temperature control in a multi-tubular ﬁxed bed Fischer-Tropsch reactor using encapsulated phase change materials
Odunsi, Ademola O.
MetadataShow full item record
The Fischer-Tropsch synthesis is a highly exothermic, indirect, catalytic, gas (syngas) liquefaction chemical process. Temperature control is particularly critical to the process in order to ensure longevity of the catalyst, optimise the product distribution, and to ensure thermo-mechanical reliability of the entire process. This thesis proposes and models the use of encapsulated, phase change material, in conjunction with a supervisory temperature control mechanism, as diluents for the catalytic, multi-tubular ﬁxed bed reactor in order to help mitigate the heat rejection challenges experienced in the process. The modelling was done using the Finite Element Analysis (FEA) software, COMSOL Multiphysics. In the main, three studies were considered in this thesis. In the ﬁrst study, a two dimensional quasi-homogeneous, reactor model, without and with the dissipation of the enthalpy of reaction into a near isothermal phase change material (silica encapsulated tin metal) heat sink, in a wall-cooled, single-tube ﬁxed bed reactor was implemented and the results were presented. The encapsulated phase change material was homogeneously mixed with the active catalyst pellets. The thermal buﬀering provided by the phase change material were found to induce up to 7% increase in selectivity towards the C5+ and a 2.5% reduction in selectivity towards CH4. Although there was a reduction in the conversion per pass of the limiting reactant and hydrocarbon productivity due to a reduction in reactor temperature, it was observed that for a unit molar reduction in the productivity of C5+, there was a corresponding 1.5 moles reduction in methane production. In the second study, a modiﬁed, one dimensional, α-model was derived which accounted for the heat sink eﬀect of the phase change material diluent. The resulting, less computationally cumbersome, yet suﬃciently accurate model was benchmarked against the more rigorous two-dimensional quasi-homogeneous model in order to check its ﬁdelity in predicting the reactor performance. As in the ﬁrst case study, a homogeneous distribution of the phase change material and active catalyst pellets was assumed. The α-model was able to approximate the reactor temperature proﬁle of the 2D-quasi-homogeneous reactor model to within 4% error, and consistently, slightly over-predicted the limiting reactant conversion by about 3%. Based on these comparisons, the α-model was deemed suﬃciently accurate to predict the reactor performance in place of the 2D model for the optimisation simulation in the third study. The third case study entailed simultaneously maximising the production of long chain hydrocarbon molecules and ensuring proper heat rejection from the reacting system, two desirable yet often conﬂicting operational requirements. The homogeneous distribution of the active catalyst pellets and the phase change material diluents was abandoned for a multi-zonal axial distribution in which, individual zones of the catalyst bed were diluted to varying extents. The best dilution and distribution “recipe” was determined using optimisation techniques and the previously derived modiﬁed α-model. The multi-zonal axial dilution of the catalyst bed brought about a marked increase (up to 19%) in the productivity of the long chain hydrocarbons, while ensuring a more judicious use of the catalyst bed in contrast to the homogeneous catalyst/phase change material arrangement in the previous two studies. The latent enthalpy of the metallic phase change material combined with its good thermal conductivity helped push the limits of the catalyst bed by increasing the conversion per pass beyond the typical 20-30% reported in literature, with less likelihood of either early catalyst deactivation or thermal unreliability of the reacting system. In the main, it was observed that the overall productivity of the desired C5+ could be enhanced by reducing the quantity of the catalyst pellets by a pre-deﬁned reactor volume. In addition, the reactor productivity beneﬁts from a highly active zone situated at the reactor entrance, immediately followed by a less reactive zone. This arrangement has the eﬀect of ramping the reaction rate (and in eﬀect the reactor temperature) early on, and this is kept in check by the less reactive zone immediately adjacent to the reactive one at the reactor entrance.