Frank R. Groves

High temperature gas desulfurization with elemental sulfur production

Preliminary results on the use of cerium oxide as a high-temperature desulfurization sorbent are presented. The primary advantage of cerium over current zinc-based sorbents is the potential to produce elemental sulfur during the regeneration phase of the process. Although CeO2 is less effective for H2S removal during sulfidation, the sulfided product, Ce2O2S, will react with SO2 to produce elemental sulfur directly. Rapid and complete regeneration is possible over the range of 500 to 700°C, and only elemental sulfur is formed. Elemental sulfur concentrations (considered as S2) as large as 20 mol% have been produced in the regeneration product. The sorbent has been subjected to ten sulfidation–regeneration cycles using a laboratory-scale fixed-bed reactor with negligible activity loss. Effectively complete conversion of CeO2 to Ce2O2S during sulfidation and subsequent regeneration to CeO2 was achieved in each cycle. A two-stage desulfurization process using CeO2 for bulk H2S removal followed by a zinc sorbent polishing step has been proposed to meet specifications of the integrated gasification combined cycle (IGCC) process. Economic comparison with a single-stage desulfurization process using zinc sorbent followed by elemental sulfur recovery using the direct sulfur recovery process (DSRP) shows that the two-stage cerium process may be less costly if the cerium sorbent is sufficiently durable.

Measurement of vapor—liquid equilibrium for binary systems containing polynuclear aromatic compounds

Abstract Vapor—liquid equilibrium (VLE) data are measured using a Stage-Mueller vapor—liquid recirculating still for ten binary systems containing polynuclear aromatic hydrocarbons at two temperatures (170 and 190°C). These data are correlated using five different thermodynamic activity coefficient models to determine the necessary binary interaction parameters. The measured vapor—liquid equilibrium data are compared with those predicted using Scatchard—Hildebrand regular solution theory. It is found that the regular solution theory predicts the bubble point pressures within an average absolute deviation of 2% and the vapor phase mole fractions within 10%. Therefore, regular solution theory can be used as a first approximation to determine the vapor—liquid equilibrium for these moderately nonideal systems.

CORRELATION OF SOLID-LIQUID AND VAPOR-LIQUID EQUILIBRIUM DATA FOR POLYNUCLEAR AROMATIC COMPOUNDS

Abstract Solid—liquid equilibrium (SLE) data were measured for five solutes in cis-decalin from ambient temperature to the melting point of the solutes. These SLE data plus our previous data for the same solutes in tetralin were used to determine experimental activity coefficients and interaction parameters using the UNIQUAC model. This set of interaction parameters gave a good representation of the experimental SLE data but were not satisfactory generally for predicting vapor—liquid equilibria for these systems. However, in an alternative approach, the solid solubility data and one vapor—liquid equilibrium point for each isotherm from previous work were used to determine another set of binary interaction parameters in the UNIQUAC model using a gradient-based search technique. The calculated values of parameters were used then to predict the vapor—liquid equilibrium for that system. It was found that the solid solubility and one vapor—liquid equilibrium data point were sufficient to predict the vapor—liquid equilibrium curve for the binary systems studied. Finally, all the solid solubility and vapor—liquid equilibrium data were combined to determine global parameters for the ten binary systems using the UNIQUAC model.

Modeling high temperature desulfurization in a fixed-bed reactor

Abstract A time-dependent, isothermal, heterogeneous model has been used to describe the high temperature reaction of H 2 S with a metal oxide sorbent in a fixed-bed reactor. Plug flow of reacting gas is assumed, and individual pellets react according to the unreacted core model. Five bench-scale desulfurization tests using two sorbent formulations with inlet H 2 S concentrations ranging from 500 to 14,100 ppmv have been modeled. One parameter, the effective diffusivity within a single pellet, was adjusted to match the experimental H 2 S breakthrough results. This made possible an independent prediction of the axial sulfur distribution at the conclusion of the test which compared quite well with experimental sulfur analyses. Experimental results from a large-scale reactor were also matched using an effective diffusivity determined from the bench-scale analysis.

ROBUSTNESS ANALYSIS OF FEEDBACK LINEARIZATION FOR PARAMETRIC AND STRUCTURAL UNCERTAINTIES

A theoretical approach to analyze robustness of feedback linearization is developed. Specifically, sufficient conditions for boundedncss and convergence of the system trajectories are found when feedback linearization based on a nominal mathematical model is applied to an uncertain real plant, which may have parametric and structural uncertainties. The developed approach does not require the restrictive matching condition, a global Lipschitz condition, nor the same equilibrium point for mathematical model and real plant for all possible model-plant mismatch. An example of an unstable chemical reaction in a CSTR demonstrates the utility of the developed approach in analyzing robustness of feedback linearization for nonlinear chemical processes.

ON THE ANALYSIS OF NONCATALYTIC GAS-SOLID KINETICS DATA HAVING WEAK TEMPERATURE DEPENDENCE

Abstract Noncatalytic gas-solid reactions should exhibit strong temperature dependence when the rate is controlled by surface kinetics. However, there are a number of examples in the literature where apparent activation energies less than 10 kcal/mol have been reported as being representative of intrinsic kinetics. This conclusion is often based on electrobalance data in which large gas Row rates were used to eliminate mass transfer resistance and the fact that fractional conversion-time results are consistent with the surface kinetics control version of a gas-solid reaction model. The oxidation of FeS was studied in an electrobalance reactor as a function of O2 mol fraction, temperature, and gas flow rate. The global rate was first-order in O2 and weakly dependent on temperature and flow rate. Data analysis used the approximate solution to the grain model. The single-resistance surface kinetics variation of the model provided good match with the conversion-time data, but the apparent activation energy was only about 7 kcal/mol. A two-resistance mass transfer-product layer diffusion variation provided equally good match with the data, and the dependence of reaction coefficients on reaction variables was in general agreement with theory.

Effect of pressure on solid-liquid equilibrium for the system carbon dioxide/n-decane/n-octacosane

Abstract Solid-liquid saturation lines (pressure vs temperature for a saturated solution of given composition) were determined for n-octacosane in mixtures of n-decane and carbon dioxide. The carbon dioxide content of the solutions ranged from 0 to 90 mol %, at pressures up to 25 MPa. The data were correlated by means of the perturbed hard sphere chain equation of state. The equation of state requires three pure component constants for each component plus three binary interaction parameters. Two of the interaction parameters were obtained from binary data. The third was adjusted to fit the ternary data. The correlation was satisfactory for carbon dioxide contents up to 80 mol %. For higher carbon dioxide levels the octacosane solubility predicted by the correlation was consistently low.

BATCH SORPTION BY LIGAND EXCHANGE: DETERMINATION OF INTRAPARTICLE DIFFUSIVITY

A mathematical model has been developed to describe batch sorption by ligand exchange. A ligand is sorbed onto an exchange sorbent from a finite bath. The sorbent particles are uniformly suspended in the fluid bath volume. The transport mechanisms are diffusion through the solid sorbent and external film diffusion. The equilibrium between the solid and fluid phase is described by a Langmuir isotherm. Two approaches are presented for representing the intraparticle mass transfer—diffusion in a pseudohomogeneous solid accompanied by sorption and diffusion through a saturated region to an unreacted shrinking core. The solid diffusion model, consisting of a partial differential equation, has been solved numerically by orthogonal collocation. The shrinking core model, a single ordinary differential equation, has been solved by conventional numerical integration. Both models successfully correlated data on diffusion of ammonia in copper-complexed carboxylic acid type ion exchange resin. The resultant sorbent dif...

APPROXIMATE OPTIMAL CLOSED LOOP CONTROL OF NONLINEAR SYSTEMS VIA PARAMETER SEARCH.

Abstract : The purpose of the paper is to present a simple straight-forward procedure for obtaining a near optimal closed loop algorithm for nonlinear system and compare it with linear design via optimal control theory. (Author)