Date of Award

Fall 2001

Document Type


Degree Name

Doctor of Philosophy in Chemical Engineering - (Ph.D.)


Chemical Engineering, Chemistry and Environmental Science

First Advisor

Robert Benedict Barat

Second Advisor

Piero M. Armenante

Third Advisor

Joseph W. Bozzelli

Fourth Advisor

Dana E. Knox

Fifth Advisor

Paul M. Lemieux

Sixth Advisor

Adel F. Sarofim


In this study, detailed thermo-chemical kinetics with networked ideal reactor model were applied to simulate a practical combustion system -the Secondary Combustion Chamber (SCC) of the Rotary Kiln incineration Simulator (RKIS) at the EPA facility at Research Triangle Park, NC. The networked ideal model was developed using analysis of reactor geometry, temperature profile measurements, and SO2 tracer data provided by EPA. A computer simulation of the networked model was developed using the CHEMKIN H library. A parallel effort considered the effects of non-ideal mixing on detailed thermo-kinetic, simulations. Specifically, an alternate approach was developed to solve the Partially Stirred Reactor (PaSR) model that allowed the incorporation of large detailed mechanisms. Both ideal and non-ideal modeling approaches were compared with experimental data gathered on a Toroidal Jet Stirred Combustor (TJSQ and the SCC at EPA. SCC experiments measured Product of Incomplete Combustion (PIC) formation of surrogate chlorinated wastes (CCl4 and CH2Cl2)lwhile the TSJC experiments measured PIC formation in ethylene/air combustion for fuel-lean conditions near blowout and fuel-rich conditions.

Analysis of the geometry and temperature profiles of the SCC suggested the existence of up to four distinct mixing zones. The RTDs, which were resolved from the tracer studies, further supported a multiple PSR model. A model was chosen based on the best fit to SO2, tracer data and consistency with physical geometry, resulting flow patterns, and temperature measurements. A thermo-kinetic mechanism developed by Chiang (1995) was applied to the model. The model results did not agree well with the experimental data. However, it followed many of the underlying trends revealed by the data. Sensitivity analysis of the parameters was used to further explore trends and recommend potential design improvements to reduce PIC formation.

An alternate solution technique was developed for the PaSR which approximated mean conditions and solved the deterministic model to refme the approximation and eventually converge on a solution. The approximation, direct integration, and convergence technique compared favorably with the published Monte Carlo modeling calculations, but used, on average, less than 1/200th of the CPU time. This new technique allowed use of considerably larger detailed mechanisms. Additionally, a generalized PaSR model was proposed to account for the effects of non-ideal macromixing.