Flow structure and heat transfer in low aspect-ratio fixed bed

Fixed-bed reactors are the most important class of reactors utilised by the chemical industry for heterogeneous catalysis. For highly exo- or endothermic reactions they are usually designed as tube and shell reactors with multiple narrow tubes to facilitate efficient heat transfer, hi such applications the tube-to-particle diameter (aspect) ratio is usually low. hi low aspect ratio fixed beds, discrete void fraction variation is more pronounced because of the presence of a close confining wall. As heat transfer is important for fixed beds applications as reactors, modelling of such systems flow structure requires a more detailed accounting of the discrete void fraction. In operation, exothermic fixed-bed reactors often exhibit regions with much higher temperature than adjacent regions and these are termed hot spots. Modelling of such a parametrically sensitive issue has so far been limited to the use of global and mean quantities of voidage, velocity and heat transfer coefficients. Most of these models have been homogeneous or pseudo homogeneous in nature, whereas few models have considered the discrete fluid flow in the bed structure. This research shows that flow structure and heat transfer patterns in low aspect ratio fixed bed reactors is better modelled by properly accounting for the discrete void fraction variations. The research initially uses illustrative 2-D models that are "randomly" packed to investigate the influence of variable local discrete voidage pattern on flow distribution. Local flow structure characterisation is by network analysis which conserves the continuity principle and energy law. Extending into 3-D using an illustrative sphere packing assembly with a low aspect ratio, discrete angular- voidage values along the axial length of the packing have been determined. The discrete flow structure and heat transfer pattern in the packing is then evaluated using the developed 3-D network-of-voids (NoV) model. The NoV model is based on network principles and accounts for the variations of discrete angular voidage variations. The NoV model offers simplicity in incorporating the discrete variations of voidage for discrete fluid flow structure characterisation. The model is not too complicated and does not require a massive computational effort and so can easily be developed to simulate a large set of randomly packed tubes in multi-tubular reactors. Wide distributions of local flow rates in a given packing have been realised. The magnitude of the discrete flow rates in the voids is largely independent of the discrete void size. This is because the discrete flow behaviour of a given void is interdependent on the behaviour of voids interconnecting it and cannot be characterised in isolation. The distribution of discrete voids velocities shows a wide variability and existence of low and high velocity gradients and in some cases reverse flows in the inter-particle voids. Phenomenal variations of discrete wall heat transfer coefficients within a single tube are encountered which imply that the different discrete sections of the tube will transfer heat at radically different rates resulting in potentially large temperature differences in different sections of the tube. These can potentially develop into localised hot spots, with several potentially unanticipated consequences for safety and integrity of the reactor. Comparison of the predicted global parameters evaluated from the discrete values predicted by the NoV model with those from models that can only predict global values shows quite reasonable agreement.