An experimental and modeling study of the dynamic behavior of an individual catalyst pellet as influenced by deactivation and dilution has been made. The reaction system employed was benzene hydrogenation over nickel/kieselguhr catalyst pellets diluted with 10% graphite, 10% silica, or undiluted; thiophene was employed as the catalyst poison.A one-dimensional effective transport model was developed, and the parameters for this model were determined principally from separate off line experiments or steady state experimental information. The adequacy of the model was determined from a direct comparison of the experimental and computed time-temperature profiles.
SCOPEThe application of simulation models to the transient response of fresh and deactivated catalyst particles, differing in physical properties, is treated. Benzene hydrogenation over a nickel/kieselgehr catalyst, poisoned by thiophene is used as an experimental example of the class of exothermic catalytic reactions subject to irreversible poisoning. The catalysts investigated included formulations diluted with silica or graphite as well as the original material and exhibited a range of physical properties.Major objectives of the study are investigations of the feasibility of completely a priori simulations (all parameters determined independently) to the extreme case of transient response represented by start-up of the fresh and deactivated catalyst, the effort required to obtain a given level of detail in simulation, and the parametric sensitivity of such catalyst reaction-poison systems.
CONCLUSIONS AND SIGNIFICANCESteady state and transient response behavior are strongly influenced by catalyst deactivation. A full distributed parameter model, described previously, was used to simulate the transient response of fresh and deactivated catalyst pellets possessing differing intraparticle transport properties. Attempts at full a priori simulation at the highest level of detail (spatial and temporal variation of intraparticle temperature profiles) were not successful. Investigation of the parametric response of the model revealed extreme sensitivity to the pellet intraparticle diffusivity and external heat transfer coefficient. Both the full model and simplified versions employing an active shell (undeactivated pellet) or a dead zone (deactivated pellet) at the surface could be used for successful simulation of the experimentally observed behavior, but only in the sense of parameter fit models.I t would appear that global, averaged quantities such as steady state activity and overall rates of deactivation can be simulated using a priori parameter sets; however, detailed simulations apparently require parameters to be known to an accuracy beyond current capabilities in estimation or measurement.In a recent paper (Lee et al., 1978), we reported on the computer simulation of a series of experiments dealing with the steady state and transient behavior of partially deactivated catalyst particles for the hydro, nenation of benzene. The simulation was attempted on ...