ABSTRACT Number: 1141152
In situ burning is an efficient response method that quickly removes large quantities of oil from the marine environment eliminating the need for collection, storage, and transport. The combustion of hydrocarbons mainly yields carbon dioxide and water; however, it also creates large plumes of black carbon soot and residues of incompletely burned oils. Three research projects focusing on improving burn efficiencies show promise to make an already efficient response method even more efficient. Specifically, a technology to increase heat transfer back into the crude oil result in more complete combustion greatly reducing carbon soot is nearing completion and will soon be ready for transfer to industry for commercialization. A study reconfiguring existing booming techniques allows more oxygen into the fire resulting in decreased soot production and cleaner burns. Finally, a fundamental study into the phenomena of fire whirls demonstrates a dramatic increase in volumes of oil burned while greatly reducing emissions. Emissions and efficiencies of the studies are compared with standard pool fires.
Fire. It was only until fairly recently that fire, one of human-kind's oldest tools, was considered as a tool to remediate oil spills. A team of international archeologists found evidence of controlled use of fire by homo sapiens and Neanderthals dating back 1 million years ago, and the Byzantines used burning oil known as “Greek Fire,” as weapons of war in 678 CE during the Arab siege of Constantinople (Cartwright, 2017). Yet it was not until 1958 on the frozen Makenzie River near Norman Wells in Canada's Northwest Territory, that in situ burning (ISB) was intentionally used to remove oil after a pipeline spill (McLeod and McLeod, 1972 as cited by Fingas, 2018). Roughly a decade later, the British military intentionally bombed the oil tanker Torrey Canyon off the coast of Great Britain when it accidently ran aground in March of 1967 (Walsh, 1968). During the 1970s and 80s, many intentional ISB tests were conducted in the United States and Canada (Fingas, 2018). Fast forward to March 1989, when the Exxon Valdez ran aground in Prince William Sound, Alaska spilling 11.2 million gallons of North Slope crude oil (USCG, 1993). A test burn was successfully conducted near Goose Island in Prince William Sound, but winds increased as a storm approached emulsifying the oil to 80% water rendering the oil unignitable (USCG, 1993).
The oil spill response community responded to Valdez spill with dozens of in situ burn-related studies conducted in the nineteen nineties and aughts. In August 1993, 25 governmental agencies from Canada and the United States coordinated an extensive, intentional spill and in situ burn of 20,400 gallons of crude oil off the coast of Newfoundland, Canada known as the Newfoundland Offshore Burn Experiment (NOBE) (API, 2004). Test burns at the USCG's Joint Maritime Testing Facility on Little Sand Island in Mobile Bay, and early fire boom tests and the development of a test standard for fire booms at BSEE's (formerly Mineral Management Services) Ohmsett test tank in Leonardo, New Jersey. In situ burning came into its own in the spring and summer of 2010 after the Deepwater Horizon lost well control on April 20. Over 400 successful burns took place during the response remediating five percent of the oil recovered by response efforts (Restore the Gulf.org).
Scores of ISB research has transpired over the last decade by industry, academia, the research community, and governmental agencies. The Interagency Coordinating Committee on Oil Pollution Research (ICCOPR) 2015–2021Oil Pollution Research and Technology Plan (R&T Plan) prioritized research into potential public health and environmental health from in situ burning including its effectiveness, impacts, planning and technology. The Bureau of Safety and Environmental Enforcement (BSEE) alone has sponsored and conducted over two dozen ISB-related studies addressing the research gaps in the R&T Plan.
Current research focuses on increasing the efficiency of in situ burning (ISB). It is important to note, that there are two types of efficiencies frequently expressed regarding ISB: burn efficiency and combustion efficiency. Burn efficiency is a volumetric measure relating the burn residue to the volume of the initial oil, whereas combustion efficiency relates the stoichiometric completeness of the conversion of hydrocarbons to its products carbon dioxide and water vapor. The goal of increasing burn efficiency is to reduce the volume of residual oil after a burn. A burn efficiency greater than 85% is considered highly efficient. Burn efficiency is mainly a function of the initial slick thickness (Buist, 1998). Combustion efficiency is a measure of the stoichiometric completeness of the in situ burn. During burn operations, most of the hydrocarbons combust yielding carbon dioxide and water vapor, as well as particulate matter and minor amounts of other gases. Figure 1 shows typical combustion by-products from burning crude oil. However, as evidenced by the black smoke plume shown in Figure 2, in situ burns of crude oil do not completely destroy all of the hydrocarbons. Of the average ten percent particulate matter, most of that is in the form of black carbon soot, agglomerations of solid carbon particles, which rise in the plume. This soot results when insufficient amounts of air are drawn into the fire. (Note that different types of crude oil are more prone to sooting then others.) Fires greater than 3 meters in diameter are oxygen starved, creating even more black carbon soot (Koseki, 2000). Soot precipitates out of the plume at a rate exponentially from the fire (Fingas, 2010). For large pool fires, a soot layer formed above the fuel surface blocks flame radiation that limits fuel evaporation (Rangwala, 2017). In other words, oxygen-starved fires produce more soot, which in turn, decreases the efficiency of the fire even more as the soot particle block the heat radiating from the flame needed to continue vaporizing the oil for combustion.
While the black plumes look quite ominous, they dissipate fairly quickly. The Newfoundland Offshore Burn Experiment (NOBE) found that emission levels measured 150 m from the burn were below levels that pose health concerns, and at distances greater than 500 m from the fire, it was difficult to detect any emissions (Fingas, 2011). These emissions do, however, present a health risk for first responders in the immediate vicinity of a burn.
This paper focuses on three BSEE research projects seeking to increase both the burning efficiency and the combustion efficiency associated with in situ burning, and methods used to quantify these improvements: 1) Advancing the Maturity of the Flame Refluxer™ Technology, 2) Linear Augmented Fire Boom Configuration to Increase Burn Efficiency and Reduce Emissions (with residue, emissions, and efficiencies measured and characterized by the U.S. Environmental Protection Agency (EPA)), and 3) Efficient Remediation of Oil Spill over Water Using Fire Whirls.
BSEE IN SITU BURN STUDIES
Flame Refluxer™ Technology: Worcester Polytechnic Institute: BSEE Project Nos. 1104/1068/1049
Flame Refluxer™ Background
As spilled oil weathers, its volatile components vaporize and can be ignited. A small percentage of the heat from the flames (3–5%) is transferred back to the oil slick, vaporizing the heavier components are able to vaporize (Buist, 1998). Thus, it is the vapors hovering above the oil that are burning, not the oil itself. Worcester Polytechnic Institute (WPI) developed a simple and robust system to collect the radiative and convective heat generated by combustion back to the fuel slick to create a heat feedback loop to increase the burning rate.
WPI's Flame Refluxer TM (FR) is a patented technology based on the simple concept of transferring heat from the flame back to the fuel using a metal wool blanket (immersed in the oil slick) and copper coils (located in the flame zone), thereby creating a feedback loop sustaining hotter, faster, and cleaner combustion. The technology is cheap, has no moving parts, and reusable. The copper coils collect radiative and convective heat generated by the combustion back to the fuel to create a feedback loop that sustains a significantly increased burning rate. The heat feedback system creates a low thermal resistance pathway between the flame and the fuel. Flame Refluxer™ Experiment and Results
WPI systematically explored the effect of immersed objects on burning efficiency (i.e., burning rate enhancement and emission reduction), through experiments performed in circular oil slicks of varying diameters (0.1 m to 1.4 m). The experiments were divided into 3 phases based on diameter: small/intermediate scale (0.1m and 0.28 m), large-scale (0.7 m) and field trial (1.4 m). Figure 3 illustrates the 1.4 meter prototype Flame Refluxer™ tested during large-scale at the Joint Maritime Testing Facility in Mobile, Alabama. The collector portion of the system captures the heat of the flames typically lost to the atmosphere. The energy collected is then transferred to the heater consisting of copper fibers sandwiched between copper mesh.
The experimental results showed that the heat generated by the combustion and directed back to the fuel significantly enhanced regression rate while reducing smoke emission. In Phase I (small/intermediate scale), the regression rate of the fuel (ANS crude oil) was enhanced by 360% with blanket and multiple coils, respectively. In Phase II (large-scale), Flame RefluxersTM (blanket and coils) enhanced the burning efficiency by 185% and reduced the CO/CO2 by 50% for a 1 cm oil slick on water layer.
Figure 4. clearly shows that the color of the smoke plume is much lighter for the blanket with 48 coils case with coils of varying heights (VH). Quantitative smoke measurements were not possible. The light color smoke indicates reduction in small black particles of carbon and more complete combustion. It should also be noted that the blanket with 48 coils (VH) case was burning ~4 times faster.
Large scale field tests were performed in the field and resulted in three major outcomes. The first outcome of field tests was the significant increase in steady state regression rate achieved with the blanket-mesh-coil system as shown in Fig. 5a. The regression rate reached a value of 17 mm/min at 48 coils (with coils of the same height) placed on circumference. This is ~ 6 times of the baseline case (3 mm/min) which equates to an additional heat flux of 132 kW/m2 transferred from the flame to the liquid fuel. The second outcome, the reduction of post burn residue using the FR burner, is shown in Fig. 5b. Only 1.8% of the oil was left on water when the blanket+48 coils (with coils of the same height) was used. The third outcome is the reduction in CO/CO2 ratio in the steady state burning regime also shown in Fig. 5b. The cleaner combustion was also observed visually.
The blanket increases the residue removal by extending the extinction time, adding nucleate boiling regions, and acting as a wick. During ISB, the natural extinction of the flame occurs because of the heat loss to the water sublayer. With addition of blanket, heat generated by the burn stored in the blanket and facilitates sustained combustion even in very thin oil slicks (2–3 mm). High thermal capacity of the blanket prevents the blanket to cool down even when it is in touch with water (heat sink). Field-tests performed in the outdoor Joint Maritime Test Facility (JMTF), Mobile Bay, Alabama, showed that the developed blanket-coil prototype was able to burn thin oil slicks (~1 cm) achieving an efficiency of 480% above baseline. Further, the high thermal capacity of the blanket sustained combustion even in very thin oil slicks (~1 mm). The extended burning time and oil wicking resulted in 1.8% post burn residue (baseline test comprising of no blanket had a post burn residue of 32%). Complete combustion of the fuel caused a reduction in the black smoke as evidenced by the CO/CO2 ratio reduction by half compared to the baseline.
Linear Augmented Boom: U.S. Army Corps of Engineers Cold Regions Research and Engineering Lab (CRREL): BSEE Project No. 1093
Linear Augmented Boom Background
The objective of the project was to achieve an enhanced and a cleaner in situ burn by changing the geometry of the burning oil slick and supplementing compressed air to the burn. Variation of Boom Aspect Ratios (length:width) and augmentation with compressed air injected into the flames were investigated to determine if these configurations and augmentations would allow better natural penetration of air into the burn zone to promote better mixing of the fuel and air to both reduce soot yields and increase radiant heat feedback to the burning slick increasing burning rates, thereby increasing the oil removal efficiency.
It is known that the rates of oil consumption and soot production are functions of the surface area of the burning slick. The crude oil burn rate increases with pool diameter until it reaches about 3 m, at which point oil consumption levels off at around 3.5 mm/min (Kosecki, 2000). The oil consumption rate is limited by the radiant heat transfer back to the burning slick.
Smoke is produced by the incomplete combustion of crude oil, which is largely because of a lack of oxygen, or the inability to supply sufficient air to the center of the fire. Large in situ oil fires draw in large amounts of air and most of this entrained air is drawn upwards by the rising column of hot combustion gases. These gases do not penetrate to the middle of the burning slick. Smoke yield is defined as the ratio of the smoke emission rate (mass/time) to the oil consumption rate (mass/time). The smoke yield increases with the pan diameter until about 3 m (e.g., from 0.055 kg smoke/kg oil burned in a 0.09-m pan to about 0.2 kg smoke/kg oil burned in a 3-m pan), but does not increase with further increases in diameter (Koseki, 2000). Large-scale in situ burns of crude oil in fire booms at sea have diameters on the order of 40 m and will burn at about 3.5 mm/min (Buist, 1998).
Linear Augmented Boom Experiment and Results
To explore the effect of boom geometry and air injection on burning efficiency, experiments were performed in equal area, rectangular oil slicks of varying aspect rations (1:1 to 9:1, length to width) and air volumes. The experiments were divided into two phases based on scale: intermediate-scale (1.2 m wide by 11 m long SL Ross wave tank), and large-scale (2.4 m wide, 14.3 m long, 2 m deep CRREL wave tank).
Using the results of the intermediate test burns, a detailed design for the full-scale linear augmented fire boom was produced. The design was based on enhancing a 50-foot section of commercially-available DESMI PyroBoom® fire boom with adjustable structural components to hold the fire boom in rectangular shapes of different aspect ratios in waves and angled compressed air nozzles and compressed air supply hoses. A 50-foot section of the modified fire boom was positioned to contain the experimental burns in the CRREL wave tank. The aspect ratio of the burn area in the modified fire boom was varied from 1:1 to 9:1, length to width. The area encompassed by the boom was kept constant at approximately 3.4 m2. Figure 6 shows the layout of the three boom rectangles with aspect ratios (L × W) of 1:1, 4:1, and 9:1 that fit into the working surface area of the CRREL wave tank (6 m × 2.4 m and the actual 9:1 boom configuration and one of the test burns inside the CREEL wave tank.
In general, the data from the large-scale experiments in the CRREL wave tank showed that all the experiments produced an oil removal efficiency of greater than 88%; there appears to be a slight increase in burn efficiency with increasing aspect ratio in calm conditions (due to better aeration of the flames); calm conditions generally result in higher burn efficiency than burning in waves, as one might expect (waves cause gentle mixing of the slick, which likely increases heat transfer through the burning oil which in turn causes earlier extinction); however, the presence of waves, regardless of air or boom ratio condition, always lowers the amount of oil weight loss. Higher boom ratios are more efficient than lower room ratios. The boom ratio effect is possibly due to more efficient air penetration into the flame zone due to the thinner combustion zone; increased air injection does not seem to significantly affect burn efficiency at this scale (the scatter in the data masks any effects); the lowest measured burn efficiency (88.8% mass removed) was measured with the 4:1 aspect ratio, in waves with the nozzles pointed at 135° (45° down towards the slick). This was potentially due to turbulent mixing of the burning oil layer by the jets of compressed air causing earlier onset of the vigorous burn phase and extinction than in the case of a quiescent burning oil layer.
Linear Augmented Boom Modified Combustion Efficiency
Higher Boom Ratios tended to result in higher oil weight loss unless waves were present. For MCET (with TC), however, higher Boom Ratios resulted in more efficient combustion under all groupings (Figure 8).
MCET calculated using gas and particle carbon averaged 0.89 and ranged from 0.85 to 0.93. Higher Boom Ratios have higher MCET values than lower Boom Ratios and higher oil weight loss, but only in the absence of waves. The highest MCET, 0.92, was observed at the highest Boom Ratio, 9:1. This is possibly due to more efficient air penetration into the flame zone due to the thinner oil slick configuration. High MCET values can still be associated with lower mass loss fractions depending on other conditions. Residue emission factors (the mass concentration per mass of residue) declined with increasing oil mass loss but no such correlation was found with MCET values.
The particle number concentration was dominated by particles < 100 nm. The particle size distributions generally did not change throughout the burn period. The run with the highest MCET value resulted in PM2.5 emission factors about three times lower than the lowest MCET runs.
Further testing at CRREL is warranted to assess whether additional compressed air would further reduce soot emission factors. Additional tests to determine how much more compressed air would be required to virtually eliminate soot would be very useful.
Efficient Remediation of Oil Spills over Water Using Fire Whirls: University of Maryland (UMD): BSEE Project No. 1094
Fire Whirl Background
Despite their importance and effective use, ISB techniques are still challenged by the airborne emissions they release (especially near shore), oil slick thickness, the degree of weathering and emulsification, and the intensity of ambient wind and waves. In general, burns over water are oxygen-starved (NOAA, 2017; S.& T.C. National Response Team, 1995; IPIECA, 2016); hence, ample black soot can be seen in almost any ISB. In fact, one of the major operational limits of current ISB practices is the after-burn emission concentrations, particularly at the downwind distance of the burn close to the populated areas. For instance, the maximum 1-hour averaged concentration of Particulate Matter of ten micrometers (PM10) must not exceed 150 μg/m3, according to ISB guidance from the National Response Team, Science and Technology Committee (S.& T.C. National Response Team, 1995). Reduction of PM production requires high efficiency at the time of burning, including reaching ample oxygen concentrations, high temperatures over the fuel surface and completeness of combustion (Fingas et al, 1995). UMD researchers believe this can be achieved using a phenomenon called the Fire Whirl. The fire whirl is one of the most dramatic structures which arises at the intersection of combustion and fluid dynamics, forming when the right conditions of wind and fire interact. The resulting intensification of combustion imposes significantly higher heat feedback to the fuel surface, increasing the rate of burning and potentially decreasing the post-combustion residuals and increases the amount of oxygen available and temperatures within the fire, reducing the production of PM within the plume.
The initial concept for this proposal arose from observations of fire whirls that occurred naturally and accidentally in previous ISBs. Figure 9 shows an example of the natural occurrence of an ISB-fire whirl taken in the Gulf of Mexico. Black smoke emitted from the initial pool-like fire formed between fire-resistant booms turned into a lighter, almost white smoke around the fire whirl. Such white-colored smoke indicates a significant reduction in soot and cleaner burning than in the surrounding pool fire.
Fire Whirl Experiments and Results
In this work, UMD studied the formation, applicability, efficiency, and integration of fire whirls with current ISB techniques, investigated from small to large scales. Potential benefits or drawbacks of the techniques were assessed in terms of associated emissions, unburned residue, and combustion efficiency (burning rate, heat flux and fuel surface temperature) (NOAA, 2017; S. & T.C. National Response Team, 1995). The physical mechanisms at different scales were studied and used to help define operational limits and best practices with a selected proof-of-concept design proposed at the end of the study. The objectives of this project were to (1) describe and characterize the structure and behavior of fire whirls over open water; (2) understand the effects and advantages of fire whirls on in-situ burning (ISB); and to understand and quantify emissions from both traditional ISBs (pool fires) and an enhanced configuration using fire whirls at small and large scales.
Measurements of gaseous emission (CO, CO2, O2) and particulate matter, the fuel burning rate, inlet velocity, temperature (gas and liquid phase), and heat flux from fires in both pool and fire whirl configurations were conducted in small (11 cm diameter), medium (20 cm and 30 cm diameter), and large-scale (70 cm diameter) tests. Experiments with both heptane and Alaska North Slope (ANS) crude oil at various scales have shown an increased burning efficiency of fire whirls, i.e., a higher burning or heat-release rate and therefore decreased burning time for fire whirls, typically on the order of double the rate or half the time compared to traditional pool fire configurations. Emissions similarly were reduced in a fire whirl configuration, especially as scales increased. Depending on the fuel type and scale, total particulate matter production was reduced as much as 4 times for fire whirls compared to pool fires, with almost all medium and large-scale configurations exhibiting at least a 2 times reduction in total particulate matter. While heptane tests had almost all fuel removed, some less-volatile components of ANS crude remained during tests, though this amount increased with increasing scales, on the order of 80%+ removal at large-scale, consistent with previous measurements on pool fires.
Temperature, velocity, and heat flux data help to describe some of the processes controlling the burning process. Fundamentally, improvement in burning efficiency is caused due to an increase in heat transfer to the fuel surface, which was corroborated with experimental measurements.
UMD conducted the large-scale experiments at Worcester Polytechnic Institute's fire safety lab. Preliminary experiments were performed to adjust the gap size for optimum fire whirl generation. A gap size of 0.45 m was found to consistently form fire whirls that were sustained for the burning duration. The diameter of the burn pan was 0.7 meters with fire whirls reaching 5.5 m high. The walls were wheeled away for pool fire tests. Figure 10 shows burning of ANS crude oil in the pool fire and fire whirl regimes at WPI.
Total Particulate Matter (TPM) emission factors from tests with 10 – 70 cm diameter pools are shown in Figure 11. Emission factors (EF) represent the mass of particles produced per the mass of fuel consumed. Figure 11a shows the results for heptane fires, indicating that the TPM emission factor for the heptane pool fire is about 18 g/kg-of-fuel at the largest scale, which is more than double of the heptane fire whirl at the largest scale (about 8 kg/kg-of-fuel). Figure11b shows the TPM EF from ANS fires, and the trends are similar to that of the heptane fires. The TPM EF from ANS pool fire is roughly double the fire whirl emission (60 g/kg-of-fuel). EFs are also shown to change with scale, at first decreasing as the heat flux and temperatures increase from larger, more radiant flames. The naturally-driven flow also increases circulation within the chamber which may aid in mixing. As scales increase to 70 cm diameter, some efficiency is lost as the larger fire experiences less efficient mixing between fuel and air; however, fire whirls continue to achieve EFs much lower than pool fires.
Throughout this work UMD has worked to further understand the structure and behavior of fire whirls over open water. A relationship between ambient circulation and the heat-release rate of the fire was shown to be linearly correlated from small (11 cm diameter) to large (70 cm diameter) scale. Collection of heat fluxes and flame temperatures has also helped to demonstrate the mechanisms by which improved burning efficiency is achieved, namely higher heat fluxes to the fuel surface and higher flame temperatures, both augmented by the fluid dynamic structure of the fire whirl and radiation within the swirling fire plume.
Emissions measurements have directly led to quantification of the advantages of fire whirls on ISB. In particular, particulate emissions are significantly reduced when burning crude oil in a fire whirl configuration vs. a traditional pool fire configuration, typically by at least half. The rate of burning is also dramatically increased, which would result in reduced operational burning times in the field.
Ultimately, new data provided in this report have helped to improve our fundamental understanding of fire whirls and helps pave the way toward developing significantly faster, cleaner and more efficient ISB techniques.
Testing to advance the technology readiness level (TRL) of the Flame Refluxer is underway. Research into storage, deployment/retrieval, as sell as density of the units in a realistic setting are being studied. Plans to test the advanced units in the Canadian Multiparter Research Initiative (CAMPRI) are underway. Boom geometries as studied in the Linear Augmented Boom project hold great promise to enhance in situ burn efficiencies. Furthermore, this method requires little investment from industry, as it only requires changing the techniques and geometries of booms already in inventory. Finally, the phenomenon of fire whirls holds great interest to the oil spill response research community. Studies should continue to further the understanding of the basic physical mechanisms of their formation and structure as a potentially more efficient way of burning oil on water and subsequently an improved ISB technique.