Collaborators: Dr. M. Lázaro (U. Cantabria, Spain), R. Webster (Bechtel)

Current pyrolysis models proposed to describe the thermal degradation of solid fuel sources and the associated production of flammable vapors represent one of the major bottlenecks in fire modeling. Pyrolysis corresponds to a range of complex and often poorly understood physical and chemical processes that occur in response to the gas-to-solid thermal loading, including: in-solid heat transfer processes; chemical processes that correspond to decomposition and/or oxidation chemical reactions and that result in the production of gas (the fuel vapors for the fire), liquids (tar) and solid residues (char, ash); drying processes, i.e., evaporation of the free and bound water that is present as moisture in many solid materials (e.g., wood) and that is possibly followed by migration of water vapor and re-condensation; liquid phase processes, e.g., melting, bubble formation and transport as occur in many solid materials that feature a liquid as an intermediate state in a gasification process (e.g., polymers); in-solid pressure dynamics as may occur as a result of phase changes and thermal expansion and that may lead to physical and geometrical changes such as crack formation and changes in the material porosity and permeability.

The general objective of this project is to develop pyrolysis models that can be used in Computational Fluids Dynamics (CFD) -based simulations of fire growth. We focus on comprehensive pyrolysis models that adopt a material science perspective, and provide a detailed description of the thermal degradation of the virgin solids. Comprehensive pyrolysis models typically include a large number of unknown parameters (e.g., material properties and parameters of the chemical reactions) and require a careful calibration phase. During the calibration phase, the pyrolysis model parameters are determined by comparisons with experimental data obtained in bench-scale flammability tests, for instance thermogravimetric analysis (TGA), differential scanning calorimetry (DSC), or fire calorimetry (using a cone calorimeter or a fire propagation apparatus). The comparison between model predictions and experimental data use error minimization algorithms based on advanced optimization techniques.

Current work is focused on simulating the pyrolysis processes occurring in polyvinyl chloride (PVC); the pyrolysis model uses: a global one-step Arrhenius-type pyrolysis reaction for charring materials; cone calorimeter data; two optimization techniques for parameter estimation (a stochastic hill-climber algorithm and a genetic algorithm); and different parameter estimation methodologies (corresponding to unrestricted or restricted searches of the parameter space).