Abstract Numerical simulations of multiphase flow using the k‐fluid CFD model described in Part I of this issue are presented for packed beds at various operating conditions. Both steady‐state and unsteady‐state (e.g., periodic operation) feed conditions were studied numerically. Predictions of the k‐fluid CFD model are comparable with the experimental data in the literature for liquid upflow in a cylindrical packed bed. In addition to the mean porosity and the longitudinally averaged radial porosity profile, the variance of the porosity distribution is needed for predicting the probability density function of the sectional flow velocity. In the trickling flow regime, the k‐fluid CFD model provides reasonable predictions of the global liquid saturation and the pressure gradient. Relevant applications of the k‐fluid CFD model are identified in quantifying the relationship between bed structure and flow distribution in various‐scale packed beds. The combined flow‐reaction modeling scheme is proposed through the “mixing‐cell” network concept, in which the k‐fluid CFD simulation can provide the information on sectional flow distribution.
Abstract A high‐pressure trickle‐bed reactor was used to achieue high productivity and selectivity for the manufacture of a key herbicide intermediate (α‐aminomethyl‐2‐furanmethanol) (amino alcohol, AA) from α‐nitromethyl‐2‐furanmethanol (nitro alcohol, NA). Raney Nickel catalysts of varying activity were prescreened for suitability in tricklebed operation. The effect of operating parameters such as reactant feed concentration, liquid mass uelocity, and temperature on the yield of amino alcohol (A) for RNi‐a are discussed. The superiority of trickle‐bed reactors ouer others such as semibatch and batch sluny systems is demonstrated. The AA yield increases with decrease in feed reactant concentration and liquid mass velocity, as well as with lowering of the operat ing temperature. A maximum product yield of 90.1% was obtained at 8.3 wt. % feed concentration of nitroalcohol (NA), while at the highest feed concentration of 40 wt. % NA, the maximum product yield was 58%. The volumetric productivity of AA was significantly higher at higher reactant feed concentrations, even though the yield was lower under these conditions. The operating temperature was also an important parameter, with a lower temperature being preferable for trickle‐bed experiments. Bed dilution with inert fines improued catalyst utilization and increased the AA yield and productivity in the laboratory‐scale trickle‐bed reactor.
A one-dimensional reactor and catalyst pellet scale flow-transport-reaction model utilizing the multicomponent Stefan−Maxwell formulation for inter- and intraphase transport is developed to simulate unsteady state operation in trickle bed reactors. The governing equations and method of solution are discussed. Results are presented for a model reaction system (hydrogenation of α-methylstyrene) under gas reactant limiting conditions, for liquid flow modulation as a test case of unsteady state operation. Model simulations predict that periodic liquid flow modulation can alter the supply of liquid and gaseous reactants to the catalyst and result in reactor performance enhancement above that achieved in steady state operation. The effects of key modulation parameters such as the total cycle period, cycle split, and liquid mass velocity are simulated, and model predictions are found to be in agreement with experimentally observed trends in the literature.
The Holub et al. (1992, 1993) phenomenological model for pressure drop and liquid holdup in trickle flow regime at atmospheric pressure was noted by Al-Dahhan and Dudukovic (1994) to systematically underpredict pressure drop at high pressure and high gas flow rates. In this study, the Holub et al. (1992, 1993) model has been extended to account for the interaction between the gas and liquid phases by incorporating the velocity and the shear slip factors between the phases. As a result, the prediction of pressure drop at the operating conditions of industrial interest (high pressure) has been improved noticeably without any significant loss in predictability of liquid holdup. The extended model and the comparison between its prediction and experimental high pressure and high gas flow rate data are presented and discussed.
Abstract Conventional strategies for discrimination of intrinsic and apparent kinetics from crushed‐ and whole‐catalyst‐pellet experimental data, respectively, do not yield satisfactory results for the reaction network in the manufacture of (α‐aminomethyl‐2‐furanmethanol) (aminoalcohol) from αnitromethyl‐2‐furanmethanol (nitroalcohol). Laboratory trickle‐bed reactor tests in the range of concentration and product yield of commercial interest are utilized to yield a reasonable set of kinetic parameters, which are otherwise unobtainable. This is accomplished by proposing a reaction network, a plausible mechanism, and optimizing the kinetic parameters based on the reactor‐model‐generated peformance data to fit experimental output concentrations of all species for the entire set of experiments. A complex reaction network for the key reactions in the system is developed based on the reaction scheme in Part I. Fitting of trickle‐bed reactor data to this model is attempted to yield an insight into the actual kinetics. The results show promise of obtaining an overall network kinetic model, even with the limited data available.