A Computational Fluid Dynamics Model to Simulate Wood Combustion

Abstract

Residential wood stoves and wood heaters are being used by nearly 3 billion people worldwide. In the United States, residential wood heaters are used by only 9 % of the households. However, according to the US EPA, they contribute to around 7 % of the cumulative PM2.5 emissions. Furthermore, firebrand generation from wood combustion is the primary mechanism for wild land fire spread. In order to predict various performance parameters of residential heaters and predict wood degradation behavior for fire safety, a comprehensive understanding of wood combustion is necessary. Wood combustion involves various coupled phases, namely, dehydration, pyrolysis, char oxidation, and gas phase combustion. These phases involve several heat transfer processes including radiation from the surrounding flames and surfaces onto the wood surface, convection with the bulk gas flow, radiation losses from the wood surface, and energy generation due to the exothermic char oxidation reactions. Along with these heat transfer processes, several mass transfer processes including production of water vapor (from dehydration), production of volatile combustion gases (from pyrolysis), consumption of surface oxygen (from char oxidation), and production of CO, CO2 (from char oxidation) are involved. These processes are interdependent and may occur simultaneously, thus making them strongly coupled. The aim of this dissertation is to formulate a comprehensive kinetics based numerical model to simulate wood combustion. To this end, a reduced three-step wood pyrolysis mechanism was developed using several microscale experiments and model-fitting algorithms. The reduced mechanism accounts for both the solid phase heat flow and the gas phase heat release. For char oxidation, a conjugate heat transfer driven UDF model was developed in ANSYS Fluent. This char oxidation model was validated against wind tunnel experiments of smoldering firebrands at various wind speeds. The pyrolysis and char oxidation models were coupled together with a gas phase methane combustion mechanism and implemented into the UDF framework. This coupled wood combustion model was validated against mesoscale cone calorimeter experiments with different sample sizes. The validated wood combustion model was implemented in a top-lit updraft (TLUD) wood stove being designed by the University of Alabama (UA). The TLUD wood stove simulations were validated against the steady-state gasification rate and thermocouple temperature data collected for different operating conditions. Using the TLUD wood stove simulations, soot and CO emissions at various locations in the combustor were also quantified. The wood combustion model formulated in this dissertation can be readily applicable to various fire safety and clean energy applications.