EXPERIMENTAL AND NUMERICAL STUDY ON THE EFFECT OF FUEL AND AIR ROTATION ON FIRE WHIRL CHARACTERISTICS

Abstract

With the increasing frequency and intensity of wildfires driven by climate change, understanding their mechanism is paramount for effective wildfire management. Various parameters, including wind conditions, topography, fuel load, the generation of firebrands, and fire whirls can significantly impact wildfire propagation, worsening its spread and intensity. Firebrands, which are embers or burning debris carried aloft by fire whirls, pose a significant threat during wildfires. Burning particles can travel far, starting new fires ahead of the main one by landing on rooftops, vegetation, or other flammable material. These secondary fires are often unpredictable and challenging to control, making it harder for firefighters to manage the overall spread of the fire. Understanding the dynamics of fire whirls and their capacity compared to free-burning fires is crucial for developing strategies to mitigate their impact and enhance wildfire management and prevention measures, particularly in fire-prone regions worldwide. This study aimed to quantify variations in various fire whirl properties, such as Heat Release Rate (HRR), Mass Burning Rate (MBR), flame temperature, radiative heat flux, flame geometries, and emission from the generated fire whirl. In addition to investigating different geometric and thermal aspects of fire whirls, the movement of hot firebrands with the fire whirl is recognized as a complex problem in the context of the predicting forest fire propagation. Forest fuel, comprising live and dead foliage, shrubs, plants, and trees near the actual fire, serves as the primary source of firebrands. Furthermore, when a whirl interacts with the burning forest fuel, the remaining unburnt or partially burnt fuel mass can rotate along with the whirl, introducing variability in fundamental fire whirl properties compared to swirling flames around a stationary fire. In view of these aspects, this study extensively analyzed and discussed two conditions: (i) when circulation is imposed around a stationary buoyant fire and (ii) when circulation is imposed around a fire with rotating fuel. An Emmons and Ying-type fire whirl generator (outer screen diameter = 0.5 m and length = 1.8 m) was used to investigate the change in fire whirl behavior with fuel rotation. Different instrumentations, including pitot tubes, differential pressure transducers, thermocouples, heat flux sensors, video camera, and flue gas analyzers, were used to capture respective properties of the generated fire whirl. In the first part of the thesis, which focuses on experimental investigations, liquid fuel (gasoline) was used to generate stable fire whirls, enabling a more precise study of fire characteristics. A total of 26 experiments with different imposed circulation and fuel rotation were conducted, covering a range of HRR 5-30 kW. The study introduces a modified semi-empirical equation H∗ f = K.( • Q∗ .Γ∗2) m based on experimental data, establishing connection between the MBR and the imposed circulation, as well as between flame height, MBR and circulation. The temperature measurement for different liquid fuel fire whirl experiments reveals that the continuous flame zone accounts for approximately 70% of the total flame height. While compared to pool fires, maximum centerline temperature increases up to 35% with the imposed circulation. Furthermore, it provides insight into understanding the independent effect of imposed circulation on emission parameters. Conversely, the quantitative data obtained from the gasoline fire whirl experiments can be used as a valuable foundation for comparison and further analysis. In the second part of the thesis, forest fuels were employed to replicate real-case fire whirl scenarios. Three distinct types of live forest fuels, namely Pinus Roxburghii, Shorea Robusta, and Grevillea Robusta, commonly referred to as Pine, Sal and Silver Oak trees in India, were collected from the Himalayan regions. These fuels were allowed to dry in the sunlight for two weeks before the experiments. To determine which fuel would be more vulnerable to forest fires, preliminary tests were performed using certain fixed variables. A variety of fuel pan dimensions, along with imposed circulation and fuel rotation, were utilized to conduct a total of 48 experiments in this part of the study. Similar to the liquid fuel experiments, different fire whirl properties were examined. For our set of experiments, HRR was in the range of 42-345 kW, which can be correlated with circulation using equation ˙Qmax = 87.42Γ1.09 and with fuel pan diameter using equation • Qmax = a1 db1 fp, respectively. The values of a1 and b1 may vary in different experiments. The estimation of HRR using the oxygen-depletion method and FTIR data was found to be consistent with the conventional HRR ( ˙Qavg), with a difference of ±13% for free-burning fires. This difference reduces to up to ±3% with imposed rotations. The study also includes a detailed analysis of the effect of imposed circulation and fuel rotation on major and minor effluents from the fire whirl. Compared to Free Burning (FB) fires, the concentration of major effluents in fire whirls were found to be 50% to 400% lesser in fire whirls. The measurements were compared with full-scale fire whirls, and our data appears to be in transition between medium-scale laboratory tests and large-scale field tests. Consequently, this dataset can be used for predicting the characteristics of fire whirl generated in actual forest fires. In the third part of the thesis, which focuses on numerical modelling, the Fire Dynamics Simulator (FDS) was employed as a CFD tool to model the fire whirl generated from gasoline. A pyrolysis model is utilized to simulate the evaporation of liquid fuel. Numerical results of the MBR, flame height, temperature profiles, velocity contours, and velocity profiles are presented. The comparison between a fire whirl (hot flow) and an air tornado (cold flow) reveals an increase of up to 40% in the centerline axial velocity due to buoyancy. The predicted MBR, and flame height were in agreement with the experimental data, having an uncertainty of ±14% and ±11%, respectively. The numerical simulations reveal that the swirling flame exhibits higher and more concentrated regions of HRR than the non-swirling flame. They contribute to enhance combustion, higher HRR, intensified heat, faster flame propagation, turbulence enhancement, complex flow patterns, and potential flame instabilii ities. The evolution of fire whirl was studied using velocity vectors across the fire whirl, and precise tangential and axial velocity profiles were predicted through numerical analysis. The measured and predicted flow fields were found to be in good agreement. Hence, it can be inferred that FDS is an efficient tool for studying various aspects of fire whirls. In the fourth and final part of the thesis, the focus shifts to modeling and simulation of moving fire whirls and their interaction within the Wildland-Urban Interface (WUI). Solid fuel, specifically pine needles, is considered the fuel bed, and detailed pyrolysis mechanisms of the solid fuel are incorporated into the combustion modeling. The movement of the fire whirl (Rate of Spread (ROS)) can be significantly influenced by factors such as fuel bed thickness, wind conditions, and topography. The ROS decreases as the fuel bed thickness increases, with the average ROS increasing by up to three times when the fuel bed thickness increases from 0.01 m to 0.05 m. It was found that, as the fuel thickness increases, the burn duration for which structures can be exposed to thermal radiation greater than 10 kW/m2 (threshold limit for structures ) also increases, and it is found to be approximately 97 s, 178 s and 270 s. The maximum radiative flux on the WUI structures can reach up to 95 kW/m2, 107 kW/m2 and 146 kW/m2 for fuel thicknesses of 0.01 m, 0.02 m, and 0.05 m, respectively. These datasets can be crucial for the management of wildfires involving a whirl.