Catalyst Deactivation - Spent Hydroprocessing Catalysts - page 86

Figure 4.24: Variation of demetallization rate with time on stream for different models [From refs
273 and 274 . Reprinted with permission].
The single pellet model proposed by Perreira et al. [275,276] was suitable for predicting the
performance of the bimodal catalyst comprising 0.12 and 0.60mL/g of micro- and macropore
volume, respectively. A series of correlations involving the change of parameters such as
distribution factor, reaction rate, pore radius, Thiele modulus, etc. with time on stream was
developed. A database on the effect of the feed origin and of the catalyst properties on HDM
was established. The results supported the advantages of bimodal catalysts compared with
unimodal catalysts. The model was suitable for selecting and designing the tailor-made
catalyst to match the type of catalyst with the feed properties. This includes the size and shape
of catalyst particles as well as the catalyst porosity.
Among other models, the concept introduced by Toulhoat et al. [277,278] involved the estimate
of the “ultimate storage capacity” of the single catalyst particle, defined as the mass of metals
accumulated in the HDM catalyst pellet, relative to the unit mass of the fresh catalyst, until the
catalyst activity became zero. There are a number of other models the validity of which was
tested using various feeds. For example, the models developed by Dautzenberg et al. [279] and
Hannerup and Jacobsen [54,280] are based on the pore-plugging by metals. The pore-plugging
model developed by Oyekunle and Hughes [281] predicted performance of the HDM catalyst
lasting 1 year providing that the catalyst possessed a suitable pore size distribution.
Modeling on reactor level requires all parameters used for modeling on a catalyst activity and
single particle scale. In addition, new parameters such as liquid holdup, bed height, bed
volume, catalyst volume, pellet size, superficial liquid velocity, etc. are necessary. For
fixed-bed reactors, most of the models assume trickle flow of the liquid and gaseous streams
co-currently from the top to the bottom through the void space between the catalyst particles.
The wetting of catalyst surface ensures that reactions occur predominantly in the liquid phase.
Moreover, the plug flow transport through the reactor, usually assumed for such models,