Date of Award


Document Type


Degree Name

Master of Science (MS)


Chemical Engineering


The Dakota Gasification Company in Beulah, North Dakota operates the only commercial coal-to-synthetic natural gas (SNG) plant in the United States, and continued research on the plant's operations is of prime importance. An ASPEN/SP model of a Lurgi coal gasifier that will predict the impact of variable input parameters, such as coal composition and reactant feed flows, on production rates and efficiency is the first step in constructing a model of the entire plant to control its operation and optimize the economic performance of producing SNG.

A model, called RGAS, was developed with the combined effort of several researchers (Benjamin, Denn, Dweck, etc.) during the Great Plains ASPEN Model Development project phase. This model contains parameters (i.e., equilibrium and kinetic constants, heat transfer coefficient to the jacket, and volatile heat capacity) that were verified to be the best established values or were optimized using a pattern search to obtain model predictions used for statistical analysis and comparison to plant data. The volatile heat capacity assumes that the energy associated with devolatilization is thermally neutral. However, pressure differential scanning calorimeter tests of five North Dakota lignite coal samples indicate this energy is exothermic.

In this study, some significant model improvements have been made. • The predicted flow of condensible volatile matter (naphtha, phenol, tar oil) was corrected to allow for the significant amount of non-condensible volatile matter produced during devolatilization. (Previously, all volatile matter was assumed to be condensible.) • The heat transfer coefficient of the reactor vessel was reduced from 170 to 105 W/m2K by matching steam production and enthalpy predictions to experimental data. • The kinetic rate constants used in the RGAS model were adjusted to provide predictions that would match plant data as closely as possible. Results provided predictions within 2% (relative) of plant data for raw gas mole fractions on a dry basis (i.e., H2, CO, C02, CH4). However, the raw gas and water mass flow rates and steam utilization predictions are still in error by as much as 8%, 10%, and 20%, respectively. • The predicted raw gas temperature was made to match the plant data by adjusting the specific heat of the volatile matter from 0.43862 + 1.5414(10'4)T to 0.17. We can justify this change by rejecting the previous assumption that the heat of devolatilization is neutral, since it is an exothermic process for many lignite coals (13). This was experimentally verified using a differential scanning calorimeter at operating conditions similar to those in a Lurgi gasifier for five representative samples of plant lignite.

We have proceeded as far as possible in our model improvement efforts without changing the model itself. The next logical step is to include a kinetic model of, and incorporate the drying of coal into, the devolatilization process. Since there is considerable overlap l^etween the gasification and devolatilization zones the evolution of volatile products will affect the concentration of gasification reactants, and thus the reaction rates. The current devolatilization model is a simple, linear, temperature dependant evolution of volatile matter and does not reflect actual volatile evolution. A kinetic model of Beulah lignite devolatilization with drying, the next phase of the Great Plains Gasifier Plant Optimization project, would be much more realistic and hopefully give better mass flow predictions.