(534c) Molecular-Level Modeling of Municipal Solid Waste Gasification

Horton, S. R. - Presenter, University of Delaware
Klein, M. T., University of Delaware
Zhang, Y., Air Products and Chemicals, Inc.
Petrocelli, F., Air Products and Chemicals, Inc
Bennett, C. A., University of Delaware

Molecular-level Modeling of Municipal Solid Waste Gasification


Scott R. Horton*, Craig A. Bennett, Michael T. Klein

Department of Chemical & Biomolecular Engineering, University of Delaware


Yu Zhang, Francis P. Petrocelli

Air Products and Chemicals, Inc.


                In 2010, there was 250 million tons of municipal solid waste (MSW) in the United States. Over half of this waste ended up in landfills which fail to recover the energy stored within MSW. The most basic waste-to-energy technology is incineration. While this process recovers energy, it also poses problems to the environment due to SOx, NOx, dioxin, and furan emissions. Plasma-arc gasification has nonstoichiometric amounts of oxygen and reduces the production of these harmful oxygenates while recovering energy from waste via syngas production. Furthermore, due to the extremely high temperatures of the plasma torch, the final by-product is a salable glass-like slag. The objective of this study is to aid process design by developing a molecular-level kinetics model of plasma-arc gasification.

            The MSW is typically described at the lumped level, i.e., biomass, food waste, and plastics. In the present work, each of these lumped components was described at the molecular-level. Biomass was modeled as cellulose, hemicellulose, and lignin. In addition to these molecules, lipids, common amino acids, and sugars were used to model food waste. The modeled plastics were polyethylene, polyethylene terephthalate, and polyvinyl chloride. These feeds were all represented at the molecular level using Flory statistics for nearly linear polymers and lattice theory for lignin.

            The feed stream was converted into a reaction network using an in-house software tool, the Interactive Network Generator (INGen). There were two major categories of reactions: devolatilization/pyrolysis and gasification. The major pyrolysis reaction families were thermal cracking, decarboxylation, and decarbonylation. For gasification, the dominant reaction type was reaction with oxygen. The generated reaction network was solved and optimized using another in-house software, the Kinetic Modeling Editor (KME).

           For integration with process simulation, the optimized KME model was integrated into ASPEN PLUS using user-defined models. In the case of gasification, the KME model contains reactors in series and rate laws employing linear-free energy relationships for reaction families. The process of KME-Aspen integration was generalized allowing for the automatic implementation of any reaction model into an ASPEN worksheet.

          The KME model could also be coupled to a computational fluid dynamics (CFD) solver and take equipment level information into account in the design. Once the products and kinetics of the pyrolysis and gasification reactions are obtained from KME, the reaction network could be converted into one or several global pyrolysis/gasification reactions by kinetic model reduction. The simplified mechanism could be easily integrated into any CFD solver, such as Ansys FLUENT. On the other hand, equipment information obtained from the CFD simulation, such as flow and temperature pattern, provides guidance for choosing reactors in ASPEN PLUS process simulation.