Skip to main content

PhD. Thesis Presentation - "The Behavior of Pitfires" by Veronica Kimmerly

May 17, 2022


Fire Protection Engineering Department

PhD Thesis Presentation

Thursday, May 19, 2022




“The Behavior of Pit Fires”

Veronica Kimmerly

WPI Fire Protection Engineering



In the past few years research has emerged that identifies open-air waste fires as a significant contributor to global greenhouse emissions and human health hazard with long-term consequences from exposure. Clearly these fires need to be addressed but the context of these fires in low-infrastructure or impoverished regions limits the potential solutions. Given that for many people open-air fires are the only way they can dispose of their waste besides landfilling, it is unlikely that these fires can be eliminated. So, if these fires can't be stopped is there a way they can be improved? Open-air waste fires are typically burned in a pit to limit fire spread and provide thermal protection. Previous research into the effects of a fire in a partially filled pit (ullage is the term for the distance between the fill level and the top of the pit) has found burning is enhanced for some ullage to diameter ratios. If altering the pit geometry and the fuel location within in the pit can improve these fires, either by reducing emissions or reducing exposure time, then it is a highly implementable solution to this hazard.

To investigate the effects of ullage on pit fire behavior, a series of small-scale and meso-scale experiments are performed in conjunction with the development of a one-zone model. The small-scale experiments (9 cm and 10 cm diameters) decoupled the burning rate from the air flow by using a methane gas burner. The first set of experiments analyzed the characteristics of a fire in a pit using an experimental set up comprising of a square burner with fuel (methane) injected into the burner at different ullages (U) with three flow rates (2, 4, and 6 g/min) to vary the heat release rate. The flame dynamics are captured using high-definition video. Air entrainment is measured using a vent hood, where the products of combustion are collected and detailed temperature measurements within the hood are used to determine the location of the steady smoke layer height. At all fuel flow rates, the flame fluctuates inside the ullage, opening up at the central portion at one time instant and closing down to a conical flame at another time instant. Based on the value of fuel flow rate and ullage value, the frequency of this flame fluctuation varies. At a fuel flow rate of 2 g/min, the flame is diffusion controlled. As the fuel flow rate is increased, convection begins to dominate. Based on the ullage and fuel flow rate, partial premixing of air and fuel occurs within the ullage and a bluish flame is observed showing highly efficient combustion behavior. In summary, for a given ullage, there is a range of fuel flow rates where the airflow into the ullage is strong enough to impart partial premixing. Thus, based on the fuel flow rate or the heat release rate, an appropriate ullage can be chosen to improve the burning performance. The air entrainment depends directly on the heat release rate and inversely on the ullage and fluctuation frequency values. Based on this an engineering correlation based on a similarity analysis is proposed and experimental and theoretical values of air entrainment are within reasonable agreement. The correlation is limited because assumptions of negligible radiation and decoupling of mass transfer are made when using a gas burner in the experiments. However, the correlation is valuable as it led to a systematic study of air entrainment effects on heat release rate in a pit.

The second set of small-scale experiments are conducted to study the effect of pit depth on flame pulsation, anchoring, oxygen penetration, and temperature in buoyant diffusion flames of methane in a cylindrical burner. Depending upon the ullage (i.e., pit depth), four stages of flame behavior are identified. Stage I exhibits stable flames anchored close to the porous plate with vertical pulsation. At higher ullages, the flame also exhibits a U-shaped flame with a certain pulsation frequency (termed as stage II). At even higher ullages, the anchoring location begins to vertically oscillate in the pit while still exhibiting both vertical and U-shape pulsation (termed stage III). At the highest ullages, oxygen cannot penetrate very far into the pit, causing the flame to anchor towards the top of the pit and only exhibit small vertical oscillations (termed Stage IV). Temperature, vertical pulsation frequency, flame height, and penetration depth determine that Stage III flames are the optimal flame stage and exhibit enhanced burning. A parameter characterization study has also been conducted to explain the diffusive and convective time scales in pit fire. This work indicates that an open-air pit fire can be designed with a certain ullage to diameter ratio that could have Stage III flame behavior and enhance the efficient burning of waste in these pits.

Then meso-scale experiments (57 cm diameter) were conducted with a condensed fuel (kerosene) where mass loss rate is coupled to the air entrainment and radiative heat flux. A constantly replenishing pool system was used to maintain a constant ullage below the fire. These experiments were conducted to determine if stage II fires are still enhanced at the meso-scale and if the trends in plume temperatures from the small-scale persist. The results confirm that stage II flames are enhanced, with comparable mass flux to that seen at 0D (pool fire) but with shorter flames, indicating reduced flame necking leading to a cylindrical shaped flame. The temperature profiles showed a similar trend to that seen at the small-scale with 0.75D having the highest plume temperature despite the shortest flames (and thus further from the TC tree). Finally, the experimental results are compared to data from a simple 1-D model and the model is shown to be accurate enough to use as a tool for designing future large scale experiments with condensed fuels.

In the future, more experiments are needed to determine the fire behavior at the large scale (D ≥ 1 m) and what ullage to diameter is optimal, to better visualize and model the flow field of the flame with PIV and advanced modelling, to further the understanding of radiative heat flux in the pit with more temperature measurements inside the pit cavity, and finally to determine the chemistry behind the reduction in CO production and if this reduction is significant enough to impact the health and environmental hazards presented by waste fires. 




Prof. Ali Rangwala, WPI Fire Protection Engineering

Prof. Albert Simeoni, Department Head, WPI Fire Protection Engineering

Prof. Michael Timko, Chemical Engineering, WPI

Prof. Jose Torero

Prof. Laureen Elgert, Department Head, WPI The Global School