Date of Award

1989

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Chemical Engineering

First Advisor

Douglas P. Harrison

Abstract

The widespread adoption of coal-fueled advanced electric power generation systems depends upon the ability to remove H$\sb2$S from the gasifier product. While H$\sb2$S can be removed by traditional wet scrubbing methods, sensible heat is lost, thereby reducing the overall system efficiency. High temperature desulfurization is now possible using newly developed metal oxide sorbents which are capable of reducing the H$\sb2$S concentration to less than 10 ppmv from a laboratory scale fixed-bed reactor using a simulated coal gas. In this study, reactor simulation models capable of describing the high temperature sulfidation and regeneration reactions of metal oxide sorbents in a fixed-bed reactor have been developed. The models consider the heterogeneous nature of the gas-solid reactions, the inherent unsteady state nature of the process, the fluid mechanics of flowing gas in the reactor, heat and mass transfer between gas and solid, heat lost to the surroundings, and the kinetics of the single pellets which comprise the bed. Only material balances are considered in the isothermal sulfidation model. Because the sorbent regeneration reaction is highly exothermic, both material and energy balances are included in the regeneration model. The unreacted core model is used to describe the noncatalytic gas-solid reactions in the individual pellets. Experimental results are compared with those predicted by the models. During sulfidation, the H$\sb2$S breakthrough curves as well as the axial sulfur loadings in the reactor following breakthrough predicted by the model agree with the experimental results if the pellet effective diffusivity is treated as a "best-fit" parameter. Gas temperature profiles in the reactor during regeneration agree with model predictions in terms of maximum temperature and time for the maximum temperature to be attained; however, the experimental temperature distribution is considerably broader than predicted. In another application, the model provides very close agreement with experimental gas temperature profiles in a catalyst regeneration test.

Pages

300

DOI

10.31390/gradschool_disstheses.4818

Share

COinS