Completed Research
Scaleup and scaledown of slurry reactors: A new methodology
Juan Ricardo Inga, PhD, 1997
(Thesis: UMI Dissertation Publishing)
A new methodology for scale-up and scale-down of slurry reactors was developed using two geometrically different reactors. This methodology proposed to maintain the relative extent of gas/liquid mass transfer and reaction kinetics similar in both reactors in order to scale-up or scale-down the process. To quantify the relative extent of these rates, a new approach was developed to determined the rate-limiting step, and the mass transfer and hydrodynamic characteristics were measured in both reactors. The first reactor was a 4-liter Zipper-Clave stirred reactor, and the second one was a 0.3 m inside diameter slurry bubble column. The mass transfer characteristics including gas solubility, mass transfer coefficient, gas holdup and bubble size were statistically obtained in the two reactors under a variety of operating conditions. The operating conditions used in the stirred reactor were 2-25 bars, 298-373 K, 6.66-20 Hz and 0-50 wt. % catalyst loading, while those used in the slurry bubble column were 1.5-8 bars, 0.05-0.25 m/s superficial gas velocity, and 0-50 wt.%. The gases employed were H2, N2, CO, CH4, and C2H4; the liquid was a commercial mixture of hexanes; and the catalyst was spray dried iron oxides with an average particle size of about 40 microns. The Central Composite Design method was used to evaluate the effects and the interactions of the variables on mass transfer characteristics. The solubility values of the non-hydrocarbon gases (H2, N2 and CO) appeared to decrease with temperature, whereas those for the hydrocarbon gases (CH4 and C2H4) showed an increase, and the following trend was observed: C*C2H4 > C*CH4 > C*CO > C*N2 > C*H2. The kLa values in both reactors were determined using the Transient Physical Gas Absorption Technique. The results obtained in the stirred reactor showed that mixing speed and the catalyst concentration were the most important variables that affect kLa values whereas in the slurry bubble column the main variables which affected kLa values were pressure and catalyst loading. Small gas bubbles and large gas holdup values were measured at higher pressures and an opposite effect was observed with increasing catalyst concentration. As the gas bubbles coalescence tendency increases, the bubbles grow decreasing their residence time and consequently the gas holdup. The new developed kinetics approach was coupled with the measured hydrodynamics and mass transfer characteristics in a computer model which was used to estimate the performances of both reactors. By keeping the relative rates of gas/liquid mass transfer and reaction kinetics similar, a balance was kept between the supply and consumption of the reactants in both reactors. This study led to the simulation of a conceptual Fischer-Tropsch slurry bubble column reactor of 7 m diameter and 30 m high with 10,000 bbl/d capacity, operating at 30 wt.% catalyst loading by a 2-liter stirred reactor operating at 5 wt.% catalyst loading, at 20 Hz mixing speed, 30 bars and 523 K.