No abstract

in Wiley InterScience (www.interscience.wiley.com).The Fischer-Tropsch synthesis (FTS) in which a syngas is converted into a wide range of paraffins, olefins, and oxygenates, has found renewed interest in the context of indirect conversion of natural gas. Slurry bubble column reactors (SBCR) rank high among the candidate reactors for FTS. Despite their simple construction, their design are still uncertain because of the fragmented understanding about FTS chemistry, the reactor fluid mechanics, heat and mass transfer, the thermodynamics, and how these phenomena are intermingled in the reactor. A multicomponent/compartment model was developed to account for a detailed hydrodynamics where upon were tied the Fischer-Tropsch and water-gas-shift reactions, the thermodynamics and thermal effects, the variable gas flow-rate because of chemical/physical contraction, and gas and slurry (re)circulation and percompartment back-mixing. A Cobased mechanistic kinetics accounting for olefin re-adsorption was used to describe paraffin and olefin formation and vapor-liquid equilibria were evaluated using a Peng-Robinson/Marano-Holder model. The model was used to analyze the effects of catalyst loading, temperature, gas velocity, water-gas shift, and gas contraction on the performance of SBCRs. The simulated behavior for commercial-scale SBCR was discussed in the light of a sensitivity analysis of the model outputs with regard to the hydrodynamic, the heat and mass transfer parameters.Assessment of the effect of the choice of turbulence model, and dispersive versus advective transport of catalyst.

Fischer-Tropsch synthesis (FTS), an exothermic reaction where hydrogen and carbon monoxide synthesis gas are converted to hydrocarbon products, has been under development since the 1930's. The interest in FTS depends on current and perceived future prices of crude oil but is increasingly viewed as an option for exploiting stranded natural gas. Other advantages of FTS hydrocarbons include the absence of sulphur, nitrogen, heavy metal contaminants, low aromatic content and the ability to produce high value middle distillates/fuels. Current interest is directed towards slurry bubble processes – comprising gas, liquid, and solid phases. Industrial slurry phase FTS reactors may range in size from 6 – 10 m in diameter and upwards of 30 m in height and include multiple internal heat transfer tubes. Such systems offer numerous advantages including high heat transfer rates, good mixing, and ease of online catalyst addition and withdrawal. However, one disadvantage is the complex hydrodynamics associated with slurry bubble columns, which make scale-up difficult. A literature review on heat transfer studies and correlations has been completed focusing on previous experimental setups, the synthesis of the key findings/parameters, and the identification of the necessary criteria required for reactor design and scale-up.The parameters having the most pronounced impact on heat transfer in slurry bubble columns and three-phase fluidized beds are the superficial gas velocity and liquid properties such as viscosity and surface tension, which significantly alter the bubble properties and the column hydrodynamics. The effect of particles is poorly understood and is a complex function of particle diameter and concentration. The experimental results and correlations reported here from the majority of studies are dependent upon the equipment and properties of the three phases studied – resulting in very limited applicability to other systems or for scale-up. Other concerns include the use of relatively low gas velocities, ambient temperature and pressure, relatively large particles, and relatively small columns employed in most studies, which are not relevant to industrial operating conditions. Furthermore, studies involving multiple internals were relatively few. Most columns were only equipped with a single tube or small heat flux probe thereby measuring only the local heat transfer and not taking into account the effect on column hydrodynamics of multiple internals. Of these studies only a few tubes were equipped with heaters (that did not run the entire tube length) and heat flux probes while the remaining probes were inactive.

CFD simulations have been carried out in a full three-dimensional, unsteady, Eulerian framework to simulate hydrodynamic/thermal coupling in a bubble column with internals. A first part of the study, dedicated to the hydrodynamic/thermal coupling in liquid single-phase flows, showed that assuming constant wall temperature on the internals constitutes a reasonable approximation in lieu of comprehensive simulations encompassing shell flow and coolant flow together. A second part dealing with the hydrodynamics of gas-liquid flows in a bubble column with internals showed that a RNG k-e turbulence model formulation accounting for gas-induced turbulence was a relevant choice. The last part used these conclusions to build a hydrodynamic/thermal coupling model of a gasliquid flow in a bubble column with internals. With a per-phase RNG k-e turbulence model and assuming constant wall temperature, it was possible to simulate heat transfer phenomena consistent with experimentally measured heat transfer coefficients.

No abstract

The metrology and the impact of various parameters and operating conditions on the bulk-to-tube heat transfer coefficients in two-phase bubble columns are investigated on a small-scale mock-up. It is shown that (1) quasi-adiabatic conditions can be reached in the column; (2) the bulk-to-tube heat transfer coefficients for each U-tube downward and upward sections may or may not differ significantly, depending on the way uncertainty of the measurements is estimated; (3) using the different measurements and uncertainty estimates for given conditions, a mean heat transfer coefficient over all tubes is estimated within ±5%. The consequences for bulk-to-tube heat transfer coefficient prediction in a larger column are discussed.La métrologie et les répercussions de divers paramètres et différentes conditions d'utilisation sur les coefficients de transfert de chaleur de la calandre aux tubes dans des colonnesà bulles en deux phases sontétudiées sur une maquetteà petiteéchelle. Il est démontré que (1) l'on peut parvenirà des conditions quasi adiabatiques dans la colonne; (2) les coefficients de transfert de chaleur de la calandre aux tubes pour les parties descendante et ascendante de chaque tube en << U >> peuvent différer considérablement ou non, selon la façon dont l'incertitude des mesures est estimée; (3) par l'utilisation des différentes mesures et estimations de l'incertitude pour des conditions données, un coefficient de transfert de chaleur moyen pour tous les tubes est estiméà ±5%. Les conséquences pour la prédiction des coefficients de transfert de chaleur de la calandre aux tubes dans une colonne plus grande sont discutées.

The influence of the catalyst type (Fe and Co) on CO and H 2 conversions, CO 2 selectivity, and the composition in Fischer-Tropsch synthesis slurry bubble column reactors was simulated for representative commercialscale units (7 m i.d. and 30 m height). A nonisothermal, core-annulus multicompartment multicomponent two-bubble class model was used to account for a relatively detailed hydrodynamics. It was coupled to comprehensive Fischer-Tropsch synthesis and water-gas-shift reactions, in addition to descriptions of thermodynamics and thermal effects, variable gas flow rate due to chemical/physical contraction, and gas and slurry backmixing and (re)circulation. Two mechanistic kinetic models with consideration of olefin readsorption were employed to describe the paraffin and olefin formation with cobalt-and iron-based catalysts, in addition to relatively large activities for CO 2 and oxygenate formation, mainly alcohols, for the latter catalyst. The influence of the temperature and superficial gas velocity on CO and H 2 conversions was more evident for a cobalt-based catalyst. For both catalysts, the space-dependent superficial gas velocity directly affected the gas-phase mean residence time, influencing in the return reactor temperature and conversions. Reliable estimation of the gas velocity due to chemical contraction was critical for conversions exceeding 50%. For both catalysts, the nonisothermal simulations reveal that, because heat removal is well managed from the heat-exchange area, the reactor operation can be considered as nearly isothermal.