Supporting Information for: Economic and Environmental Benefits of Higher-Octane Gasoline

Transcription

1 Supporting Information for: Economic and Environmental Benefits of Higher-Octane Gasoline Raymond L. Speth Eric W. Chow Robert Malina Steven R. H. Barrett John B. Heywood William H. Green Contents Supporting information for this paper consists of the following: This document (23 pages, 22 tables, 1 figure): Section S1: Software Descriptions Section S2: Refinery Model Parameters Section S3: Additional Sensitivity Scenarios Section S4: Detailed Sensitivity Results Section S5: Refinery Model Output Details A separate ZIP archive, containing: octane-scenarios.xls: Refinery simulation input spreadsheet solution-baseline-reference.html: PIMS output file for the reference case of the baseline scenario solution-baseline-high-premium.html: Sample output file for the high-premium case of the baseline scenario S1 Software Descriptions S1.1 GT Power GT-Power (version 7.2) is a program developed and distributed by Gamma Technologies, which models the time-dependent gas dynamics and thermodynamics of internal combustion engines. It models the induction process, imposes a time dependent heat release rate to simulate the combustion process, and models the exhaust process. GT-Power features an object-oriented structure with a large engine component library, from which the simulated engine can be assembled. We use GT Power to simulate part-load conditions for engines with different compression ratios in order to determine the relationship between compression ratio and brake efficiency. S1

2 S1.2 Aspen PIMS Aspen PIMS (Process Industry Modeling System) is a program which uses a model of an oil refinery or other petrochemical plant to determine optimal process conditions. The PIMS model of a refinery consists of models for each process unit in the refinery, along with streams representing the condition of the raw materials entering the refinery, the various intermediate process streams, and the final products. Each stream is described in terms of a flow rate, relevant physical properties (e.g. specific gravity, sulfur content) and various surrogate properties (e.g. octane number). The model for each process unit uses a linearized relationship to compute the flow rates and properties of the output streams based on the input stream flow rates and properties and process control parameters (e.g. process temperature, residence time). For each product, constraints can be placed on the properties, e.g. minimum RON, minimum and maximum specific gravity, maximum sulfur. Given a particular refinery model and prices for all purchased inputs and marketable products, PIMS solves the optimization problem of maximizing the refinery s profit by adjusting the process parameters and stream allocations to produce a high-value product slate. S2 Refinery Model Parameters Calculations for the refinery LP model were done using Aspen PIMS Version (part of aspenone V8.2). The Gulf Coast model is available to Aspen customers from http: //support.aspentech.com as Solution ID (14 May 2012). The documentation describes the model as follows: The Gulf Coast model represents a typical full-conversion refinery situated in the US Gulf Coast. It has a full complement of conversion units - FCC, Alky Unit, Coker, Hydrocracker and two catalytic reformers. The refinery produces low sulfur gasoline and distillate and is equipped with a Hydrogen Plant and several Hydrotreating Units to achieve the ultra-low sulfur constraints. The Diesel HDT can desulfurize down to 8 ppm Sulfur, and ULS Diesel sulfur specification is max at 15 ppm. Gasoline sulfur is max at 30 ppm - a Gas Oil Hydrotreater is available to desulfurize FCC feed, and there is an FCC Gasoline HDT (Scanfiner) to further FCC gasoline sulfur. There are two physical and three logical crude units, the last one producing lube base stocks. The first crude unit processes low sulfur crudes, and the atmospheric resid can go to the FCCU. The second logical unit processes high-sulfur crudes with the vacuum residue feeding the Coker. [Nomenclature used above: FCC (Fluid Catalytic Cracker); HDT (Hydrotreater); ULS (Ultra-Low Sulfur); FCCU (Fluid Catalytic Cracking Unit).] Gasoline produced by the refinery is subject to the following property constraints in the model: RON and AKI: scenario and grade dependent Ethanol: scenario dependent S2

3 Reid Vapor Pressure: 9.0 psi specific gravity: > 0.70 Surrogates for distillation curve requirements: Percent evaporated at 160 F: 15 50% Percent evaporated at 200 F: 35 70% Percent evaporated at 300 F: > 70% Aromatics: < 55% Benzenes: < 2% Olefins: < 25% Sulfur: < 30 ppm wt. Reformate: < 50% Alkylate: < 30% Modeled volatile organics, toxics, NOx, and exhaust benzenes The following modifications were made to the Aspen Gulf Coast model to create the baseline scenario. All prices are in 2011 U.S. dollars. Total distillation capacity set to 100,000 barrels per day Fixed crude slate, to reflect inability of the aggregate system to change crude slate Allow desulfurization of additional distillate streams KE1 (straight-run kerosene) and LCO (cat. cracked diesel) to reduce production of residual oil set capacity constraints HCU, distillate hydrocracker: 9,000 barrels per day IS4, C4 isomerization: 4,000 barrels per day LPR, catalytic reformer: 21,000 barrels per day SFA, sulfuric acid alkylation unit: 7,000 barrels per day DLC, delayed coker: 17,000 barrels per day Disable certain atypical process units DIM, the propylene dimerization unit AT3 and VT3, distillation units for lubricant production ARU, the aromatics reformer S3

4 Adjust octane properties of certain refinery streams (as recommended by BP) More severe octane loss in Scanfiner ( SGTU ). Decreases both RON and MON by 3 points. Decrease octane of coker naphtha ( LKT ) to 65 RON, 62 MON. Set blending octane numbers for ethanol based on ethanol composition and target gasoline RON and MON. See S2.1 for details. Decrease octane of C4-derived alkylate to 94.5 RON, 91.3 MON. Adjust gasoline specifications 10% ethanol in both grades Maximum 30% alkylate in each grade Increase maximum percent evaporated at 160 F to 50%, in accordance with ASTM rules for ethanol blends. See ASTM D4814, Table 1, Note G. Increase maximum weight percent oxygen to accommodate arbitrary levels of ethanol blending. Set minimum RON to 98 for premium and 92 for regular. Remove minimum AKI for both grades. Pricing (2011 dollars) based on EIA (Energy Information Administration) projections for the year 2040 Crude oil: Consumption-weighted average of domestic and imported crudes, using WTI (West Texas Intermediate) as a surrogate for domestically produced crude. Domestic: 6.13 million barrels per day at $ per barrel. Imported: 7.57 million barrels per day at $ per barrel. Weighted average: $ per barrel. Ethanol: EIA estimate of wholesale price is $2.48 per gallon. Gasoline: EIA estimate of wholesale price is $3.80 per gallon. Take this value as representing a pool consisting of 10% premium and 90% regular. Let the price of premium be 4% higher than the price of regular. This gives prices of $3.784 per gallon for regular and $3.936 for premium. Diesel: EIA wholesale price estimate is $4.38 per gallon. Jet fuel. EIA wholesale estimate is $4.11 per gallon. We increase this price by 6% in order to get reasonable jet fuel production from the model. high sulfur (> 1% sulfur) fuel oil: Effectively disable production by setting price to $0.47 per gallon. low sulfur (< 1% sulfur) fuel oil: Use EIA retail price estimate for residual fuel oil ($3.58 per gallon), less 5% for distribution costs. Straight-run naphtha: Set price equal to LPG (liquified petroleum gas) on an energy basis. S4

5 LPG (propane): Use EIA retail price estimate of $2.00 per gallon, less 5% for distribution costs. Butane: Set price 10% higher than propane. Petroleum Coke: Set price on an energy-equivalent basis with coal, $9.19 per BOE (barrel of oil equivalent). Excess refinery fuel / still gas: Set price as half that of natural gas on an energyequivalent basis, $6.39 per 1000 standard cubic feet. Natural gas: Use price for natural gas delivered for industrial use, $9.09 per 1000 standard cubic feet. Electricity: EIA price for industrial use of $0.078 per kwh. A Microsoft Excel spreadsheet defining these modifications, along with all of the refinery parameter sensitivity scenarios, is also included with the supporting information. This spreadsheet defines a PIMS CASE table which can be applied to the Gulf Coast model available from Aspen. A summary HTML output file for the baseline scenario is included to allow verification of the PIMS solution. S2.1 Ethanol Blending Octane Numbers Recent literature has shown that the octane numbers of ethanol/hydrocarbon blends are better described in terms of the molar composition of the blend rather than its volumetric composition 1 4. Since the calculation of octane numbers in PIMS is done on a volumetric basis, obtaining the desired behavior requires determining an equivalent volumetric blending octane number for ethanol that depends on both the ethanol content and octane specification of the blend. Following Anderson et al., 3 we assume that the octane number (RON or MON) for ethanol/hydrocarbon mixtures follow a quadratic relationship in the mole fraction of ethanol x e : ON g = x e ON e + (1 x e ) ON h + P x e (1 x e )(ON e ON h ) where ON h, ON e, and ON g are the octane numbers of the hydrocarbon blendstock, ethanol, and blended gasoline, respectively, and P is the quadratic fitting parameter. Anderson et. al found P = for RON and P = for MON. However, other studies have observed less non-linear relationships for ethanol/hydrocarbon blends. We therefore choose intermediate estimates of P = 0.23 for RON and P = 0.54 for MON. The volumetric blending octane number for ethanol ON v is determined according to ON g = v e ON v + (1 x e ) ON h where v e is the volume fraction of ethanol in the blend. Given a target blend octane rating ON g and volume fraction of ethanol, these two equations can be solved to find the appropriate value for ON v. To convert between volume fraction and mole fraction, we take the density of the hydrocarbon blendstock to be 0.72 kg/l and the molecular weight to be 100 g/mol. For example, for a 10% ethanol blend with a target RON of 98, this method gives a blending RON for ethanol of 129. S5

6 CO 2 emissions reduction [Mt/year] CO 2 emissions reduction Direct monetary savings Combined monetized value Net savings, [billion USD (2011) per year] Figure S1: Reduction in CO 2 emissions, direct monetary savings due to changes in fuel production and consumption, and net societal benefit (combining direct monetary savings with monetized value of CO 2 emissions reduction based on the social cost of carbon) associated with switching to high usage of premium gasoline. Sensitivities of these benefits to changes in refinery configuration, and to fuel specifications using the advanced refinery configuration. Numerical data used in this figure are tabulated in tables S1 S3. S3 Additional Sensitivity Scenarios S3.1 Individual Process Unit Capacities In this section, we consider refinery configurations where individual process unit capacities have been increased to eliminate bottlenecks. The combination of these capacity expansions is the advanced refinery. For each modified refinery configuration, we examine the differences in product slates, costs, and refinery emissions between the reference and high-premium cases. The total U.S. impact for these scenarios is summarized in Figure S1b. When the capacity of the alkylation unit is expanded by 28%, actual utilization increases by 5% and gasoline production in the high-premium case increases by 1.9% compared to S6

7 the baseline scenario. The reduction in CO 2 emissions associated with transitioning to high-ron gasoline increases by 5.2 Mt/y, while the economic benefit associated with the transition decreases by $0.26 billion per year. Increasing the capacity of the hydrocracker by 33% increases the production of gasoline by 1.6% in the reference case and 2.2% in the high-premium case. Production of light naphtha in the high-premium case decreases 52% compared to the baseline, with a corresponding relative increase in production of diesel and jet fuel. Keeping more of the product slate in higher-value products leads to a $0.6 billion per year increase in the economic benefit associated with the high-premium case. However, there is also a 1.6 Mt/y decrease in the emissions reduction. Increasing the delayed coker capacity by 18% mainly increases the production of diesel and low-ron gasoline blending components. In the reference case, these low-ron components increase gasoline production. In the high-premium case, they displace other low-ron components from the gasoline blending pool, resulting in a $0.5 billion per year reduction in the direct savings. The addition of a propylene dimerization unit allows the refinery to convert a stream that was previously useful only as refinery fuel into a relatively high-ron (95 RON) gasoline blending component 5. This increases gasoline production by 6.1% in the reference case, and decreases the drop in gasoline production in the high-premium case by 7%. Adding this process unit has significant benefits in the high-premium case, increasing the annual cost savings by $1.5 billion and the CO 2 emissions reduction by Mt/y compared to the baseline scenario. Adding the propylene dimerization process eliminates excess fuel gas production; the refinery purchases natural gas to satisfy its net fuel demand. S3.2 Advanced Refinery Sensitivities We also considered sensitivity scenarios based on the advanced refinery configuration, the results of which are summarized in Figure S1c. The advanced refinery has more flexibility to increase gasoline production with increasing gasoline price than the standard refinery. In the high-premium case, a 10% increase in gasoline price results in a 3.7% increase in gasoline production, compared to only a 0.6% increase for the standard refinery. This is accompanied by the elimination of the light naphtha product stream, leading to a $ billion increase in the economic benefit and an additional 4.0 Mt/y reduction in CO 2 emissions compared to the advanced refinery with baseline product prices. The combination of the advanced refinery and E15 exhibits better performance in the high-premium case than the standard refinery with E15. The reduction in CO 2 emissions is greater by Mt/y and the annual cost savings is $0.2 billion greater than with the standard refinery. Increasing the octane standard for premium to 100 RON results in a similar CO 2 emissions reduction as the 98 RON premium scenario, but the economic benefit is significantly reduced. S7

8 S4 Detailed Sensitivity Results S4.1 Summary Results The following tables provide numerical values for the cases shown in Figure 4 in the main paper, as well as for several additional sensitivity scenarios described elsewhere in this document. Direct CO 2 Social Net Monetary Emissions cost of Societal Savings Reduction carbon Benefit Scenario B$/y Mt/y B$/y B$/y Baseline Low naphtha price High gasoline price High diesel+jet price High alkylation capacity High hydrocracker capacity High coker capacity Add dimerization unit Advanced refinery RON with E RON with E RON with E RON with E RON with E AKI with E Adv. refinery with high gasoline price Adv. refinery with E Adv. refinery with 100 RON and E Table S1: For CR/ON = 0.17, the total change in direct monetary savings, CO 2 emissions, social cost of arbon, and combined societal benefit between low premium and high premium cases in each scenario. S8

9 Direct CO 2 Social Net Monetary Emissions cost of Societal Savings Reduction carbon Benefit Scenario B$/y Mt/y B$/y B$/y Baseline Low naphtha price High gasoline price High diesel+jet price High alkylation capacity High hydrocracker capacity High coker capacity Add dimerization unit Advanced refinery RON with E RON with E RON with E RON with E RON with E AKI with E Adv. refinery with high gasoline price Adv. refinery with E Adv. refinery with 100 RON and E Table S2: For CR/ON = 0.21, the total change in direct monetary savings, CO 2 emissions, social cost of arbon, and combined societal benefit between low premium and high premium cases in each scenario. S9

10 Direct CO 2 Social Net Monetary Emissions cost of Societal Savings Reduction carbon Benefit Scenario B$/y Mt/y B$/y B$/y Baseline Low naphtha price High gasoline price High diesel+jet price High alkylation capacity High hydrocracker capacity High coker capacity Add dimerization unit Advanced refinery RON with E RON with E RON with E RON with E RON with E AKI with E Adv. refinery with high gasoline price Adv. refinery with E Adv. refinery with 100 RON and E Table S3: For CR/ON = 0.25, the total change in direct monetary savings, CO 2 emissions, social cost of arbon, and combined societal benefit between low premium and high premium cases in each scenario S4.2 Detailed Results Each of the following tables shows how the refinery product slate changes when going from 10% premium gasoline to 80% premium gasoline production for a particular scenario, consisting of the refinery configuration, fuel specifications, and product prices. Each row shows how one component of the refinery contributes to the overall change in the CO 2 emissions and cost to the system consisting of the refinery and the consumers of the refinery s products. The Refinery row shows the input crude oil consumption rate (100,000 barrels per day), and the change in emissions from the refinery itself, as well as the lifecycle emissions associated with refinery inputs such as electricity and natural gas. The Gasoline row combines the reduction in fuel consumption by the LDV (lightduty vehicle) fleet with the change in gasoline production at the refinery. The difference between these changes is made up by importing or exporting gasoline from the system. The CO 2 emissions change reflects the reduced combustion emissions (assuming a constant fleet distance traveled), combined with the lifecycle emissions associated with the production of the net gasoline import/export. The cost associated with gasoline is the cost of purchasing the imported gasoline (which will be negative if excess gasoline is being exported). The range of values between the high CR/ON and low CR/ON cases reflect the variability S10

11 Refinery % Gasoline % Ethanol % Diesel % Jet Fuel % LPG % Fuel Oil % Light Naphtha Coke (BOE) % Fuel Gas (BOE) % Total % Table S4: Changes in refinery product slate and emissions associated with producing 80% premium gasoline for the Baseline scenario: Standard refinery and prices, 92 RON regular, 98 RON premium, E10. in this parameter. The Ethanol row reflects the reduction in ethanol consumption, which is proportional to the reduction in gasoline consumption. The reduction in CO 2 emissions reduction and cost correspond to the lifecycle emissions and price of ethanol. For the remaining refinery products, an increase in refinery production leads to a net export, leading to an economic benefit proportional to the market price of that product. The net CO 2 change for each product is a combination of the refining emissions associated with that net export, plus the difference in combustion emissions between that product and the product it is assumed to displace on the market, if different (e.g. light naphtha may displace LPG or natural gas). The total change in CO 2 emissions and economic impact are used to compute the the impact per unit volume of gasoline production. S11

12 S4.2.1 Sensitivity to product prices Refinery % Gasoline % Ethanol % Diesel % Jet Fuel % LPG % Fuel Oil % Light Naphtha Coke (BOE) % Fuel Gas (BOE) % Total % Table S5: Changes in refinery product slate and emissions associated with producing 80% premium gasoline for the Low naphtha price scenario: Same specifications as the baseline scenario, except with the price of light straight run at half that of LPG on an energyequivalent basis. Refinery % Gasoline % Ethanol % Diesel % Jet Fuel % LPG % Fuel Oil % Light Naphtha Coke (BOE) % Fuel Gas (BOE) % Total % Table S6: Changes in refinery product slate and emissions associated with producing 80% premium gasoline for the high gasoline price scenario: Same specifications as the baseline scenario, except with the price of gasoline (both grades) increased by 10%. S12

13 Refinery % Gasoline % Ethanol % Diesel % Jet Fuel % LPG % Fuel Oil % Light Naphtha Coke (BOE) % Fuel Gas (BOE) % Total % Table S7: Changes in refinery product slate and emissions associated with producing 80% premium gasoline for the high diesel + jet price scenario: Same specifications as the baseline scenario, except with the price of diesel and jet fuel increased by 10%. S13

14 S4.2.2 Sensitivity to refinery unit capacities Refinery % Gasoline % Ethanol % Diesel % Jet Fuel % LPG % Fuel Oil % Light Naphtha Coke (BOE) % Fuel Gas (BOE) % Total % Table S8: Changes in refinery product slate and emissions associated with producing 80% premium gasoline for the high alkylation capacity scenario: Same specifications as the baseline scenario, except with the capacity of the alkylation unit increased by 28%. Refinery % Gasoline % Ethanol % Diesel % Jet Fuel % LPG % Fuel Oil % Light Naphtha Coke (BOE) % Fuel Gas (BOE) % Total % Table S9: Changes in refinery product slate and emissions associated with producing 80% premium gasoline for the high hydrocracker capacity scenario: Same specifications as the baseline scenario, except with the capacity of the hydrocracker increased by 33%. S14

15 Refinery % Gasoline % Ethanol % Diesel % Jet Fuel % LPG % Fuel Oil % Light Naphtha Coke (BOE) % Fuel Gas (BOE) % Total % Table S10: Changes in refinery product slate and emissions associated with producing 80% premium gasoline for the high coker capacity scenario: Same specifications as the baseline scenario, except with the capacity of the delayed coker increased by 18%. Refinery % Gasoline % Ethanol % Diesel % Jet Fuel % LPG % Fuel Oil % Light Naphtha Coke (BOE) % Total % Table S11: Changes in refinery product slate and emissions associated with producing 80% premium gasoline for the add dimerization unit scenario: Same specifications as the baseline scenario, except with the addition of a propylene dimerization unit to the refinery. S15

16 Refinery % Gasoline % Ethanol % Diesel % Jet Fuel % LPG % Fuel Oil % Coke (BOE) % Total % Table S12: Changes in refinery product slate and emissions associated with producing 80% premium gasoline for the advanced refinery scenario: Same specifications as the baseline scenario, except with capacity expansions for the alkylation unit, hydrocracker, and delayed coker, and the addition of a propylene dimerization unit to the refinery. S16

17 S4.2.3 Sensitivity to ethanol content and octane specification Refinery % Gasoline % Ethanol % Diesel % Jet Fuel % LPG % Fuel Oil % Coke (BOE) % Fuel Gas (BOE) % Total % Table S13: Changes in refinery product slate and emissions associated with producing 80% premium gasoline for the 98 RON with E15 scenario: Same specifications as the baseline scenario, but with 15% ethanol blending (E15) in both grades of gasoline. Refinery % Gasoline % Ethanol % Diesel % Jet Fuel % LPG % Fuel Oil % Coke (BOE) % Fuel Gas (BOE) % Butane % Total % Table S14: Changes in refinery product slate and emissions associated with producing 80% premium gasoline for the 98 RON with E20 scenario: Same specifications as the baseline scenario, but with 20% ethanol blending (E20) in both grades of gasoline. S17

18 Refinery % Gasoline % Ethanol % Diesel % Jet Fuel % LPG % Fuel Oil % Light Naphtha Coke (BOE) % Fuel Gas (BOE) % Total % Table S15: Changes in refinery product slate and emissions associated with producing 80% premium gasoline for the 100 RON with E10 scenario: Same specifications as the baseline scenario, but with an octane specification of 100 RON for premium gasoline. Refinery % Gasoline % Ethanol % Diesel % Jet Fuel % LPG % Fuel Oil % Light Naphtha Coke (BOE) % Fuel Gas (BOE) % Total % Table S16: Changes in refinery product slate and emissions associated with producing 80% premium gasoline for the 100 RON with E15 scenario: Same specifications as the baseline scenario, but with an octane specification of 100 RON for premium gasoline and 15% ethanol blending in both grades of gasoline. S18

19 Refinery % Gasoline % Ethanol % Diesel % Jet Fuel % LPG % Fuel Oil % Coke (BOE) % Fuel Gas (BOE) % Butane % Total % Table S17: Changes in refinery product slate and emissions associated with producing 80% premium gasoline for the 100 RON with E20 scenario: Same specifications as the baseline scenario, but with an octane specification of 100 RON for premium gasoline and 20% ethanol blending in both grades of gasoline. Refinery % Gasoline % Ethanol % Diesel % Jet Fuel % LPG % Fuel Oil % Light Naphtha Coke (BOE) % Fuel Gas (BOE) % Total % Table S18: Changes in refinery product slate and emissions associated with producing 80% premium gasoline for the 93 AKI with E10 scenario: Same specifications as the baseline scenario, but with an octane specification of 87 AKI for regular gasoline and 93 AKI for premium gasoline. S19

20 S4.2.4 Advanced refinery scenarios Refinery % Gasoline % Ethanol % Diesel % Jet Fuel % LPG % Fuel Oil % Coke (BOE) % Total % Table S19: Changes in refinery product slate and emissions associated with producing 80% premium gasoline for the advanced refinery + high gasoline price scenario: Same specifications as the advanced refinery scenario, but with a 10% increase in the price of both grades of gasoline. Refinery % Gasoline % Ethanol % Diesel % Jet Fuel % LPG % Fuel Oil % Coke (BOE) % Total % Table S20: Changes in refinery product slate and emissions associated with producing 80% premium gasoline for the advanced refinery with E15 scenario: Same specifications as the advanced refinery scenario, but with 15% ethanol blending in both grades of gasoline. S20

21 Refinery % Gasoline % Ethanol % Diesel % Jet Fuel % LPG % Fuel Oil % Light Naphtha Coke (BOE) % Total % Table S21: Changes in refinery product slate and emissions associated with producing 80% premium gasoline for the advanced refinery with 100 RON and E15 scenario: Same specifications as the advanced refinery scenario, but with 15% ethanol blending in both grades of gasoline and an octane specification of 100 RON for premium gasoline. S21

22 S5 Refinery Model Output Details Premium Regular Premium Scenario Fraction RON MON AKI RON MON AKI Baseline 10% Baseline 80% Low naphtha price 10% Low naphtha price 80% High gasoline price 10% High gasoline price 80% High diesel+jet price 10% High diesel+jet price 80% High alkylation capacity 10% High alkylation capacity 80% High hydrocracker capacity 10% High hydrocracker capacity 80% High coker capacity 10% High coker capacity 80% Add dimerization unit 10% Add dimerization unit 80% Advanced refinery 10% Advanced refinery 80% RON with E15 10% RON with E15 80% RON with E20 10% RON with E20 80% RON with E10 10% RON with E10 80% RON with E15 10% RON with E15 80% RON with E20 10% RON with E20 80% AKI with E10 10% AKI with E10 80% Adv. ref. w/ high gasoline price 10% Adv. ref. w/ high gasoline price 80% Adv. ref. w/ E15 10% Adv. ref. w/ E15 80% Adv. ref. w/ 100 RON and E15 10% Adv. ref. w/ 100 RON and E15 80% Table S22: Octane ratings for regular and premium gasoline, for both low-premium and high-premium cases in each scenario. S22

23 References [1] Determination of the potential property ranges of mid-level ethanol blends. Technical report, American Petroleum Institute, April [2] J. E. Anderson, U. Kramer, S. A. Mueller, and T. J. Wallington. Octane numbers of ethanol and methanol gasoline blends estimated from molar concentrations. Energy & Fuels, 24: , [3] J. E. Anderson, T. G. Leone, M. H. Shelby, T. J. Wallington, J. J. Bizub, M. Foster, M. G. Lynskey, and D. Polovina. Octane numbers of ethanol-gasoline blends: Measurements and novel estimation method from molar composition. SAE Technical Paper , April [4] T. M. Foong, K. J. Morganti, M. J. Brear, G. da Silva, Y. Yang, and F. L. Dryer. The octane numbers of ethanol blended with gasoline and its surrogates. Fuel, 115(0): , [5] J. H. Gary and G. E. Handwerk. Petroleum Refining: Technology and Economics. Marcel Dekker, Inc., fourth edition, S23