This dissertation presents the derivation, numerical implementation, and verification/validation of a generalized model that can be used to simulate the pyrolysis, gasification, and burning of a wide range of solid fuels encountered in fires. The model can be applied to noncharring and charring solids, composites, intumescent coatings, and smolder in porous media. Care is taken to make the model as general as possible, allowing the user to determine the appropriate level of complexity to include in a simulation. The model considers a user–specified number of gas phase and condensed phase species, each having its own temperature–dependent thermophysical properties. Any number of heterogeneous (gas–solid) or homogeneous (solid–solid or gas-gas) reactions can be specified. Both in–depth radiation transfer through semi–transparent media and radiation transport across pores are considered. Volume change (surface regression or swelling/intumescence) is handled by allowing the size of grid points to change as dictated by mass conservation. All volatiles generated inside the solid escape to the ambient with no resistance to mass transfer unless a pressure solver is invoked; the resultant flow of volatiles is then calculated according to Darcy’s law. A gas phase convective–diffusive solver can be invoked to determine the composition of the volatiles. Oxidative pyrolysis is simulated by modeling diffusion of oxygen from the ambient into the pyrolyzing solid where it may participate in reactions. Consequently, the mass flux and composition of volatiles escaping from the solid can be calculated. To aid in determining the required input parameters, the model is coupled to a genetic algorithm that can be used to estimate the required input parameters from bench–scale fire tests or thermogravimetric analysis.
Standalone model predictions are compared to experimental data for the thermo– oxidative decomposition of non–charring and charring solids, as well as the gasification and swelling of an intumescent coating and forward smolder propagation in polyurethane foam. Genetic algorithm optimization is used to extract the required input parameters from the experimental data, and the optimized model calculations agree well with the experimental data. Blind simulations indicate that the predictive capabilities of the model are generally good, particularly considering the complexity of the problems simulated.