Combustion, Heat Transfer and Soot Formation in Biomass-Burning Cookstoves

Thesis (Ph.D.)--University of Washington, 2019Over three billion people in the world rely on biomass for their cooking and heating needs. Indoor cooking using biomass has been identified as a significant contributor to cancer, respiratory illnesses and cardiovascular ailments in humans, resulting in over four million premature deaths every year. Improved biomass cookstoves that are more energy efficient and emit less particulate matter may help mitigate this public health crisis. This work examines combustion, fluid flow, and heat transfer in biomass-burning cookstoves, and soot formation in turbulent ethylene flames impinging on a cookpot, a simplified configuration that allows study of particulate matter behavior during the interaction of a flame with a cold surface. The goal of this work is to develop generalizable results and methodologies for understanding and predicting the performance of biomass-burning cookstoves. A two-dimensional steady-state axisymmetric CFD model of a natural-draft, biomass-burning cookstove is developed. The model includes coupled sub-models representing combustion, turbulence, and heat transfer. The model is validated against experimental data and used to predict temperatures and flow inside the cookstove, including the airflow rate through the cookstove and heat transfer to the cookpot. Excess air is found to be typically many times stoichiometric air during standard operating conditions and is sensitive to flow field obstructions. The effects of geometric and operational features such as the pot support height, secondary air entrainment, cone-deck shape, and baffle placement within the cookstove on the flow, airflow rate, mixing, and stove thermal efficiency are analyzed. The model shows that secondary air entrainment, though ineffective by itself, increases turbulent mixing when used in conjunction with a central baffle but reduces thermal efficiencies due to enhanced heat transfer to the walls. Thirty-six cone-deck configurations are modeled, and it is found that the cone-deck shape primarily affects the airflow rate through the stove, with more constricted designs leading to higher thermal efficiencies. The modeling shows that merely restricting the airflow is not a sufficient condition for increasing the thermal efficiency; different constrictions resulted in different thermal efficiencies. A similar mathematical model is applied to three variations of a three-dimensional geometry representing a real cookstove to study the effects of secondary air entrainment. The results show that a choke ring just upstream of the riser leads to relatively well-mixed flow in the riser. Secondary air from cut-outs in the front of the stove of all three configurations is found to be ineffective for entraining air into the combustion chamber but is effective for cooling the front of the combustion chamber, possibly increasing the durability of the cookstove. Soot formation is investigated experimentally in turbulent non-premixed ethylene flames impinging on a cold surface, which is a combustion configuration closely related to the biomass-burning cookstove. The cold surface is a stainless-steel receptacle containing water maintained at the boiling point (373 K). The jet diameter, jet velocity, and the distance of the surface the flame impinges from the jet nozzle is varied. Soot mass emissions are measured using a tapered element oscillating microbalance (TEOM) and size distributions using a scanning mobility particle sizer (SMPS). For free flames, both the soot yield factor and the mean aerodynamic diameter are strongly correlated with Richardson’s ratio and the jet exit strain rate. The particle size distribution of soot emitted from free flames is found to be bimodal at low jet exit strain rates and tends towards unimodal distributions as the strain rate is increased. When impinging on the cold surface, it is observed that for most flames the soot yield factor initially increases with non-dimensional surface height h* (defined as the surface height divided by the flame length), reaches a maximum when the surface is at approximately half the flame height, and then decreases. The particle size distributions of emitted soot for nearly all configurations are found to be bimodal. A two-dimensional axisymmetric CFD model is developed in order to gain insight into the soot formation process and to understand and interpret the experimental results obtained. Combustion is modeled by the non-premixed, non-adiabatic flamelet method and soot formation modeled by the Method of Moments by Interpolative Closure, with oxidation by O2 modeled by the method developed by Khosousi et al., based on the thermal age of soot particles. The model’s predictions are compared with in-flame soot and temperature data from literature. The model correctly predicts the trend of increasing, then decreasing soot mass emissions from a flame impinging on a pot, but consistently under-predicts soot mass emissions. The model is also found to be sensitive to a parameter in the formulation of the thermal age-based oxidation model, Ta,max, the value of the thermal age at the point of maximum soot volume fraction. In order to understand the increase in soot mass emissions when a cold surface is placed in the flame, the environment experienced by soot particles in a representative flame with and without a cold surface is explored. The path traced by a particle exhibiting maximum soot in both configurations is considered. Analyzing the oxidation and surface growth along these pathlines shows that both processes are essentially concluded by a residence time of 30 ms. Surface growth for both configurations reaches a maximum at approximately the same residence time but falls much quicker for the free flame. Oxidation for the free flame reaches a higher peak value, and the absence of the pot results in greater oxidation. These two effects of greater soot surface growth and reduced soot oxidation lead to increased soot emission when a pot is present as opposed to a free flame. The increasing, then decreasing trend of soot emission with increasing pot height is investigated by analyzing the soot surface growth term for a representative flame. A lower surface height causes surface growth to be disrupted, leading to a lower mass of soot generated and emitted