COMPARATIVE FIRE RISK OF MOTOR VEHICLE FUELS: GASOLINE VS ETHANOL

Transcription

1 Paper 4B COMPARATIVE FIRE RISK OF MOTOR VEHICLE FUELS: GASOLINE VS ETHANOL S.E. Dillon, A.R. Carpenter and R.A. Ogle Exponent, Wood Dale, IL, USA Prepared for Presentation at American Institute of Chemical Engineers 2008 Spring National Meeting 42 nd Annual Loss Prevention Symposium New Orleans, LA April 7-9, 2008 UNPUBLISHED AIChE shall not be responsible for statements or opinions contained in papers or printed in its publications 311

2 COMPARATIVE FIRE RISK OF MOTOR VEHICLE FUELS: GASOLINE VS ETHANOL S.E. Dillon, A.R. Carpenter and R.A. Ogle Exponent, Wood Dale, IL, USA ABSTRACT If motor vehicles switch from gasoline to E-85 ethanol, is the fire risk to the general public reduced or increased? This paper explores the fire risk associated with four accident scenarios: Transport from refinery to distributor (pipeline versus truck or barge), Transport from distributor to retail facility by truck, Spill during fuel dispensing to motor vehicle, and Motor vehicle fire following traffic accident. Each of these scenarios represents a stage in the life cycle of the liquid fuel, from wholesale distribution and consumer usage. In each case the risk of fire is composed of the probability of occurrence and the consequences of the fire. The probabilities were estimated from various compilations of accident statistics. Overall, the probability of an E-85 ethanol release was greater than or equal to a gasoline release. This was due to the comparative accident frequency of different transportation modes and distances involved. The fire consequences were quantified in terms of spill area, heat release rate and radiant heat flux to nearby targets. For spills of the same size, the gasoline fire tended to be more severe than the E-85 ethanol fire. This is due to the differences in their physical properties, the enthalpies of combustion, and the emissivities of the flames. INTRODUCTION Gasoline blended with ethanol (fuel ethanol) is becoming increasing popular as an alternative fuel for motor vehicles. As more motor vehicles switch from gasoline to fuel ethanol, is the fire risk to the general public reduced or increased? This paper explores the fire risk associated with four accident scenarios: Transport from refinery to distributor (pipeline versus truck or barge), Transport from distributor to retail facility by truck, Spill during fuel dispensing to motor vehicle, and Motor vehicle fire following traffic accident. 312

3 Each of these scenarios represents a stage in the life cycle of the liquid fuel, from wholesale distribution to consumer usage. In each case the risk of fire is composed of the probability of occurrence and the consequences of the fire. The probabilities were estimated from various compilations of accident statistics. Overall, the probability of an ethanol release was greater than or equal to a gasoline release. This was due to the comparative accident frequency of different transportation modes and distances involved. The fire consequences were quantified in terms of spill area, heat release rate and radiant heat flux to nearby targets. For spills of the same size, the gasoline fire tended to be more severe than the ethanol fire. This is due to the differences in their physical properties, the enthalpies of combustion, and the emissivities of the flames. BACKGROUND The Energy Policy Act (EPAct) was passed in 1992 and amended in 2005 to enhance United States energy security and to improve air quality. Under EPAct, the U.S. Department of Energy (DOE) hopes to decrease U.S. dependence on foreign petroleum and increase energy security by supporting the domestic production of alternative fuels. Several parts of EPAct encourage the use of alternative fuels in motor vehicles, including blends of ethanol and gasoline. In December 2007, the Energy Independence and Security Act was signed and expanded the 2005 Renewable Fuels Standard (RFS) requirements to 36 billion gallons of renewable fuel to be used annually by [1] The amount of conventional biofuel (ethanol derived from corn starch) will be required to increase from 9 billion gallons in 2008 to 15 billion gallons by 2015 (see Figure 1) Total Renewable Fuel Conventional Biofuel (Ethanol) 30 Billions of Gallons Year 313

4 Figure 1. Required volumes of renewable fuel to be sold or introduced into commerce in the United States annually. [1] Currently, one third of the gasoline used in the United States is 10% ethanol (E10) to comply with oxygenate requirements for federal clean air programs, to increase octane, and to extend the U.S. petroleum fuel supply. Another ethanol fuel that is gaining popularity in the U.S. is E85 (85% ethanol, 15% gasoline by volume). [1] The primary motivations for consumers to purchase flexible fuel vehicles 1 and alternative fuel are the increased environmental benefits and the reduced dependence on foreign fossil fuels. Ethanol is produced chemically from ethylene or biologically from grains, agricultural waste, or other materials containing starch or sugar. When produced biologically, ethanol represents a renewable fuel source. In the U.S., a majority of ethanol is produced from corn. Pure ethanol is produced at an ethanol production facility and is denatured with a small volume of gasoline (2 to 5%) to make it unfit for human consumption. Denatured ethanol is transported to fuel suppliers via railcar, barge, and truck, where it is stored. The fuel supplier typically mixes 85% denatured ethanol with 15% gasoline into a tanker truck that is then delivered to retailers for distribution to consumers. Similar to gasoline and diesel fuels, E85 is seasonally blended to aid in engine starting and provide proper performance in different geographical locations. Up to 30% gasoline is blended with ethanol (E70, winter blend) to increase the vapor pressure for starting the engine in cold temperatures. [3] Compared with gasoline-fueled vehicles, E85-fueled vehicles produce lower carbon monoxide (CO) and carbon dioxide (CO 2 ) emissions, the same or lower amounts of hydrocarbons (HC) and non-methane hydrocarbon emissions (NMHC), and a similar amount of nitrogen oxide (NOx) emissions. E85 also has a lower vapor pressure than gasoline resulting in reduced pollutants due to evaporation of the liquid fuel into the atmosphere. E85 has approximately 27% less energy than an equivalent amount of gasoline and therefore gets approximately 72% of the fuel efficiency, as measured in miles per gallon. [1] Material properties of gasoline, ethanol, and E85 are provided in Table 1. As many properties of E85 are not available in the literature, effective properties based on 15% gasoline and 85% ethanol by volume were calculated and are provided for comparative reference. Additional testing of E85 is necessary to provide a more direct comparison between fuels. Table 1. Comparison of the Properties of Gasoline and E85. [1-8] Property Gasoline* Ethanol E85** Molecular Weight g/mol g/mol g/mol (a) Octane Gallon Equivalent to Gasoline Miles per Gallon Compared to Gasoline % 70% 72% Boiling Point F F F 1 Vehicles that can run on both traditional gasoline and various blends of ethanol fuels. 314

5 Table 1. Comparison of the Properties of Gasoline and E85. [1-8] Property Gasoline* Ethanol E85** Flash Point (Closed Cup) F F F Auto-ignition Temperature F F 495 F Lower Explosive Limit (LEL) % % % Upper Explosive Limit (UEL) % 19.0% 19.0 % Latent Heat of Vaporization kj/kg kj/kg kj/g (a) Heat of Combustion kj/g kj/g kj/g (a) Lower Heating Value 18,000 19,000 Btu/lb 11,500 11,585 Btu/lb 12,500 Btu/lb Radiative Fraction (b) 40% 20% 25% Stochiometric Air/Fuel Ratio (by weight) Density kg/m³ kg/m³ kg/m³ (a) Density kg/m³ kg/m³ kg/m³ (a) 60 F Specific Heat C kg/m³ lb/gal 2.22 kj/kg K Btu/lb F 792 kg/m³ 6.61 lb/gal 2.43 kj/kg K 2.40 kj/kg K (a) Specific 60 F Viscosity, 68 F Viscosity, 68 F Reid Vapor F Vapor Pressure mm²/s 1.52 mm²/s 0.29 cp 1.2 cp 7 15 psi 2.3 psi 8 15 psi psi Vapor Density Water Solubility g/l g/l *Properties of gasoline can range depending on blend, octane, and test method used to determine the property. ** Properties of E85 will vary depending on the actual concentration of gasoline added (the blend), and the test method used. (a) Approximated based on properties of gasoline (15%) and Ethanol (85%) (b) Estimated values. Although there are some restrictions on the components that can be used with ethanol blended fuels due to material incompatibility issues, the same types of dispensing methods that are used to dispense gasoline are used to dispense E

6 The number of E85 fueling stations in the US is increasing rapidly. In 2003 there were approximately 200 stations; in September 2007 the number surpassed 1,200 and continues to increase. Based on the data available at the time of the authoring of this paper, there are approximately 1,450 E85 fueling stations in the U.S. The majority of these stations are located in the Midwest, with the highest concentrations being in Minnesota, Illinois, Wisconsin, Missouri, and Indiana. Figure 2 depicts the national distribution of E85 fueling stations. [7] Figure 2. E85 Fueling Location in the U.S. (September 2007) [3]. In comparison, as of October 2006 there were approximately 167,000 retail gasoline outlets in the United States. [14] The total annual volume of U.S. petroleum products supplied is shown in Table 2. In 2005, total U.S. demand for petroleum was 20.8 million barrels per day, of which 12.5 million barrels per day, or 60 percent, was from net imports. [14] This equates to a demand of approximately 7.59 billion barrels of gasoline per year or approximately 417 billion gallons consumed per year. A record 9.16 million barrels per day were consumed in 2005, which equates to approximately 504 million gallons in a single day. 316

7 Table 2. U.S. Finished Motor Gasoline Supplied, Including All Ethanol Blended Gasoline (E10 and E85). 2 [14] * Million Barrels 3,101 3,143 3,229 3,261 3,333 3,343 3,377 3,392 per year Billion Gallons per year * Estimate based on January through October monthly values. TECHNICAL APPROACH This paper seeks to evaluate the impact on transportation fuel fire risk as the use of E85 grows. To accomplish this goal, specific accident scenarios are developed, the fire hazards for each scenario are evaluated, and the probability of fire occurrences is estimated. To that end, four separate risk scenarios are developed which track the lifecycle of motor vehicle fuel: transport from manufacturer to distributor, transportation to retailer, spill by consumer, and motor vehicle (fuel fed) fires. The consequence of accidents occurring are then evaluated by comparing the probable spill volumes, heat release rate, and flame height that would be anticipated for both gasoline and E85. As a method of evaluating the relative risk associated between gasoline and E85 for each of the scenarios discussed, the number of fires involving motor vehicle fuel is estimated by normalizing the frequency of spill occurrence. An overriding assumption in this analysis is that spills that result in fires from gasoline, will also result in fires from E85. For the given scenarios, the combustion characteristics for gasoline and E85 do not change, however the probable spill size does. The available accident statistics vary considerably from one scenario to another; this presents some difficulty in comparing different scenarios due to differences in quality and quantity of the accident data available. Each scenario contains a brief description of the accident, the relative fire hazards of gasoline compared to E85, and the comparative probability of a spill. SCENARIO 1: TRANSPORT TO DISTRIBUTOR Scenario 1 is based on the transportation of the liquid fuel from the manufacturing facility (refinery or fermentation plant) to the distribution facility (i.e., the tank farm ). The most common transportation mode for gasoline is by pipeline. The most common mode for ethanol (actually denatured ethanol) is by tanker truck. Due in part to water s high affinity for ethanol, the corrosiveness of ethanol with typical pipeline materials of construction, and the location of existing pipelines, ethanol, denatured ethanol, and E85, are not amenable for pipeline transport and are generally transported by truck. The probability of a spill from a pipeline is considerably less than from a tanker truck. Based on a report by Allegro, a medium-sized spill from a pipeline is approximately 3,150 (11,920 liters). [15] Although rail transport would present an 2 Finished motor gasoline supplied approximately represents consumption; computed as field production, plus refinery production, plus imports, plus unaccounted for crude oil, minus stock change, minus crude oil losses, minus refinery inputs, minus exports. 3 1 Barrel = 42 US gallons 317

8 intermediate probability of spill between pipeline and truck, many distribution facilities are not equipped for tank car receiving and unloading. Typical petroleum tankers hold between 6,500 to 9,500 gallons in three to four bulkhead compartments, with the largest compartment holding approximately 3,000 gallons 4. For this analysis, a spill volume of 3,000 gallons (11,360 liters) from a single ruptured fuel compartment will be analyzed. In determining the primary hazards associated with a liquid hydrocarbon pool fire (heat release rate, flame height, and heat exposure), the overall area of the liquid pool becomes the most important variable. As the area of the pool fire grows, the heat release rate and the flame height also grow. Calculated heat release rate and flame heights for gasoline, ethanol, and E85 fuels are presented in Figure 3 and Figure 4. Note that for the analysis used, data is not readily available for E85. Therefore, the property data used to calculate the heat release rate and flame height for E85 were estimated using the data available for ethanol and gasoline and the known volumes of each fuel, 85% and 15% respectively Ethanol Gasoline E85 Heat Release Rate (MW) E85 Gasoline Ethanol Pool Fire Area (ft²) Figure 3. Calculated Heat Release Rates for Gasoline, Ethanol, and E85 Pool Fires. [10] 4 Compartment number and volume vary among tanker manufactures and specific tanker design 318

9 Ethanol Gasoline E85 Gasoline 35 Flame Height (ft) E Ethanol Pool Fire Area (ft²) Figure 4. Calculated Flame Height for Gasoline, Ethanol, and E85 Pool Fires. [10] A recommended design fire for designing fire protection systems for road tunnels assumes a fire from a gasoline tanker measuring 323 to 1076 ft² (30 to 100 m²) or a pool diameter measuring 20 to 37 ft (6.2 to 11.3 m). [9] For this analysis a pool area of 1076 ft² (37-ft diameter) was chosen as the fire associated with a tank truck spill. For a volume of 3,000 gallons, this would correspond to a pool of fuel that is 4.5 inches (11.4 centimeters) deep. Liquid pool fires resulting from spills of hydrocarbon fuels liquid are typically on the order of approximately 0.05 inches (1 millimeter) deep on smooth, unfinished concrete. Even assuming that the liquid spill in this analysis occurs on a rough surface, a pool depth of no more than a few millimeters would be expected. For this analysis, it is assumed that the spill is both bounded by nearby obstructions (curbs, variations in elevation, etc.) and that the spill does not occur all at once, allowing fuel to pour into the pool from the ruptured tank over time. In regards to the exposure from the resulting pool fire, an exposure heat flux of 5.0 kw/m² is being used. At this flux level a person would receive second-degree burns in approximately 40 seconds. It is anticipated that a person nearby would be able to move from the area to safety in less than 40 second. In order to calculate the radiative flux to a target, a radiative fraction must be determined for the particular fuel that is burning. The radiative fraction is the percentage of the total heat release from the fire that is being released in the form of radiant energy, as opposed to convective energy. Hydrocarbon fuels that produce a lot of soot, such as gasoline, typically have a radiative fraction on the order of 0.35 to 0.40 (35% to 40% of the energy is released as radiation). For clean burning fuels like ethanol, the radiative fraction is closer to For this analysis, the radiative fractions will be taken as 0.35 for gasoline, 0.20 for ethanol, and 0.25 for E85 to account for the soot production from the gasoline. 319

10 The key risk analysis parameters are the following: Gasoline Pipeline Spill size: liquid spill of 3,150 gallons (11,920 liters) over an 1,076 ft² (100m²) area Fire hazard of spill: Heat Release Rate = 240 MW Flame Height 5 = ft (17 22 m) Burning Duration = 27 minutes Distance to Safe Heat Flux, 5 kw/m² = 120 ft (36.5 m), measured from the center of the pool fire. Normalized frequency of spill occurrence: 1.0 [15, p. 36] Denatured Ethanol Tanker Truck Spill size = 3,000 gallons over 1,076 ft² Fire hazard of spill: Heat Release Rate = 40 MW Flame Height = ft ( m) Burning Duration = 100 minutes Distance to Safe Heat Flux, 5 kw/m² = 37 ft (11.3 m) Normalized frequency of spill occurrence: 46.0 (relative to pipeline spills) [15, p. 36] Thus, for Scenario 1, the probability of a spill is greater with tanker truck transport than with pipeline transport. However, since the spill size is the same, E85 poses the smaller fire hazard. Given that E85 represents a small percentage of the total volume of transportation liquid fuels, the incremental risk increase is not significant. SCENARIO 2: TRANSPORT TO RETAILER Scenario 2 is based on the transportation of the liquid fuel from the distributor to the retailer (i.e., the service station ). In most cases, all motor vehicle fuels including gasoline and E85 are transported from the distribution site to the retailer via tanker truck. The accident scenario considered here involves damage occurring to a tanker trailer (as a result of collision, or other event) that results in the loss of containment of one compartment in the tanker. This accident or damage can occur at any point along the transportation route, either at the distribution hub, at the service station, or over the road. For this portion of the analysis, a smaller spill area is used than in Scenario 1. For this scenario, a spill area of 500 ft² (46.5 m²) with a pool diameter of approximately 25 ft (7.6 m) is used. This will require the pool to have a larger overall depth and therefore allow the pool to burn longer. The choice of this pool area is somewhat arbitrary, however, the results are used to show the difference in hazard between the different fuels. The commodity within the tanker truck, whether gasoline or E85, does not influence the probability of a spill. Thus, the normalized frequency of spill occurrence is the same for both commodities. The key risk analysis parameters are the following: 5 Flame heights calculated using the methods of both Heskested and Thomas. Flame height values are shown as a range. 320

11 Gasoline Tanker Truck Spill size = 3,000 gallons over 500 ft² Fire hazard of spill: Heat Release Rate = 112 MW Flame Height = ft ( m) Burning Duration = 55 minutes Distance to Safe Heat Flux, 5 kw/m² = 82 ft (25 m) Normalized frequency of spill occurrence: 1.0 E85 Tanker Truck Spill size = 3,000 gallons over 500 ft² Fire hazard of spill: Heat Release Rate = 29 MW Flame Height = ft ( m) Burning Duration = 152 minutes Distance to Safe Heat Flux, 5 kw/m² = 35 ft (10.7 m) Normalized frequency of spill occurrence: 1.0 (relative to gasoline tanker spills) Thus, for Scenario 2, the probability of a spill is the same for E85 as it is for gasoline. Since the spill size is the same, E85 poses the smaller fire hazard. SCENARIO 3: CONSUMER SPILL Scenario 3 is based on the refueling of a motor vehicle by a consumer at a retail fueling station. This accident scenario considers either spilling resulting from over filling of the vehicle tank (auto shutoff mechanism missing, inoperative, or defeated), over filling of a portable container, or spilling as a result of human error. Small fuel spills of this nature likely occur on a regular basis at refilling stations worldwide but go unreported or unnoticed if the resulting spilled fuel does not ignite. However, the potential for a fire to occur exists for all of these spills and the likelihood of ignition increases as the volume of spilled fuel increases. According to the National Fire Protection Association (NFPA) vehicle fire data, an average of 4,100 highway vehicle fires (1% of the total) occurred at service or gas stations. [11] A recommended design fire for designing fire protection systems for road tunnels assumes a fire from a passenger car measuring 22 ft² (2 m²). Based on an estimated spill thickness of approximately 1/8 inch (3 mm), this corresponds to a fuel spill of approximately 1.6 gallons (6 liters). In this analysis, it is assumed that all spills are ignited. Since the liquid fuel delivery mechanism is the same for both gasoline and E85, the normalized frequency of spill occurrence is assigned the same value. The key risk analysis parameters are the following: Gasoline Spill Spill size: liquid spill of 1.6 gallons (6 liters) over a 22 ft² (2.0 m²) area Fire hazard of spill: Heat Release Rate = 4.6 MW Flame Height = ft ( m) 321

12 Burning Duration = 40 seconds Distance to Safe Heat Flux, 5 kw/m² = 17 ft (5.2 m) Normalized frequency of spill occurrence: 1.0 E85 Spill Spill size: liquid spill of 1.6 gallons over a 22 ft² area Fire hazard of spill: Heat Release Rate = 1.2 MW Flame Height = 8 ft (2.4 m) Burning Duration = 112 seconds Distance to Safe Heat Flux, 5 kw/m² = 8.5 ft (2.6 m) Normalized frequency of spill occurrence: 1.0 (relative to gasoline spills) Thus, for Scenario 3, the probability of a spill is the same for E85 as it is for gasoline. Since the spill size is the same, E85 poses the smaller fire hazard. SCENARIO 4: MOTOR VEHICLE ACCIDENT Scenario 4 is based on the inventory of liquid fuel in a motor vehicle. This scenario assumes that one or more vehicles participate in a collision that results in the loss of containment of the fuel tank for one of the vehicles. Automobile accidents occur frequently, however automobile accidents that result in loss of fuel containment are only a small subset of total vehicle accidents. This is due to numerous factors including the design and performance of vehicle safety systems, the locations of fuel tanks within the vehicle, and the nature and dynamics of specific automobile accidents. There are, however, a significant number of vehicle fires that occur that are considered fuel fed that involve the vehicle fuel. As a method of simplifying this analysis, the authors assume that an average of 10 gallons (38 liters) of fuel participate in a typical fuel fed vehicle fire. This is a necessary assumption to make due to the large variety of fuel tank sizes on different makes and models of vehicles and the varying amount of fuel that can be present at any point between refueling. To account for natural changes in elevation, and allowing for a rough ground surface and assuming a non-instantaneous fuel spill, a fuel depth of approximately 1/4 inch (6 mm) will be used. This corresponds to a fuel spill area of 70 ft² (6.5 m²) with a diameter of 9.4 ft (2.9 m). The probability of a liquid fuel spill does not depend on the contents of the fuel tank. Thus, the normalized frequency of spill occurrence is assumed to be the same for gasoline and E85. The key risk analysis parameters are the following: Gasoline-Fueled Vehicle Spill size: 10 gallons (38 liters) over a 70 ft² (6.5 m²) area Fire hazard of spill: Heat Release Rate = 15.6 MW Flame Height = ft ( m) Burning Duration = 79 seconds Distance to Safe Heat Flux, 5 kw/m² = 31 ft (9.5 m) Normalized frequency of spill occurrence:

13 E85-Fueled Vehicle Spill size: 10 gallons over a 70 ft² area Fire hazard of spill: Heat Release Rate = 4.0 MW Flame Height = 12 ft (3.6 m) Burning Duration = 219 seconds Distance to Safe Heat Flux, 5 kw/m² = 13 ft (4.0 m) Normalized frequency of spill occurrence: 1.0 (relative to gasoline-fueled vehicles) Thus, for Scenario 4, the probability of a spill is the same for E85 as it is for gasoline. Since the spill size is the same, E85 poses the smaller fire hazard. CONCLUSIONS Of the four accident scenarios considered, only Scenario 1 (Transport to Distributor) would likely lead to a greater incidence of liquid fuel spills. The risk of fire due to accidental spills of E85 is offset by its reduced fire hazard compared to gasoline. A further consideration is that the total expected production volume of E85 is small compared to gasoline. Thus, the incremental fire risk posed by E85 is judged to be negligible compared to gasoline. In the other three scenarios, the fire risk of E85 is lower than for gasoline due to E85 s lower heat release rate. On balance, the substitution of E85 for gasoline poses slight reduction in fire risk to the public. REFERENCES 1. Renewable Fuels Association website, 2. Handbook for Handling, Storing, and Dispensing E85, U.S. Department of Energy, July U.S. Department of Energy - Energy Efficiency and Renewable Energy, Alternative Fuels and Advanced Vehicles Data Center, 4. The SFPE Handbook of Fire Protection Engineering, 3 rd Edition, Society of Fire Protection Engineers, Bethesda, MD, Fire Protection Handbook, 19 th Edition, National Fire Protection Association, Quincy, MA, Quintiere, J. G., Fundamentals of Fire Phenomena, John Wiley & Sons, LTD, Methanol Institute website, 8. CountryMark E-85 Ethanol Fuel Blend Material Safety Data Sheet, 9. Arthur G. Bendelius, Fire Protection for Road Tunnels, Fire Protection Handbook, 19 th Edition, Chapter 7, Section 14, National Fire Protection Association, Quincy, MA, 2003, p Babrauskas, V., Heat Release Rate, Fire Protection Handbook, 19 th Edition, Chapter 3, Section 1, National Fire Protection Association, Quincy, MA, Marty Ahrens, U.S. Vehicle Fire Trends and Patterns, National Fire Protection Association, Quincy, MA, October

14 12. Warner, C.Y., James, M.B., and Woolley, R.L., A Perspective on Automobile Crash Fires, Society of Automotive Engineers Report No , National Ethanol Vehicle Coalition website, Energy Information Administration website, Trench, C., The U.S. Oil Pipeline Industry s Safety Performance, Allegro Energy Consulting, (February 2003). 324