Ethanol is widely employed as an oxygenate, octane enhancer and renewable resource for gasoline blending. It is important to demonstrate comparative emissions effects of ethanol blending on air quality, through studies of both evaporative and tailpipe emissions from automobiles and light-duty trucks. A team from Future Fuel Strategies and THiggins Energy Consulting comprised of Dr. Nigel Clark, David McKain, Terry Higgins and me set out to develop a guide, a pathway, for characterizing the emissions in a representative fashion, using blends that are expected in the marketplace at present (E10) and increased (E15, E20) ethanol blending levels. The final study is available and a webinar recorded to explain the results. Those materials are available here:
Ethanol blending with gasoline yields nonlinear property changes. Gasoline is a complex mix of petroleum components, and a real-world mix is needed to reflect the nonlinear behavior and provide accurate emissions prediction. Composition of the petroleum fraction is governed strongly by refinery economics. Customarily the concentration of aromatics in the gasoline is reduced by the refiner in response to increasing ethanol content, because the ethanol offers a high blending octane number and reduces the petroleum blending octane requirement.
This study or guide presents a step by step approach to designing a comparative market fuel emissions study. This should matter greatly to the affected industries — auto, oil, ethanol — because it is these studies that are being used to design fuel/emissions regulations that directly impact the industry. The guide emphasizes the importance of aromatic reduction, and the changes in gasoline distillation that are due to ethanol blend effects and are that are expected to be seen at the pump. Prior evaporative emissions studies have not employed representative fuel composition, neglecting the sensitivity of fuel system components to the fuel makeup.
Multivariate studies of gasoline composition effects on tailpipe emissions have employed a matrix of fuels, typically adjusting properties in a way that does not reflect market fuel metrics. Future studies should characterize fuels based on detailed hydrocarbon analysis and composition, rather than standardized tests that do not correspond to the fuel behavior in the engine. Doing so is expected to provide a more accurate and realistic characterization of ethanol’s impact on vehicle emissions. The guide also supports the use of representative driving cycles and presents changing automobile engine technology that may alter comparative emissions and demand new studies.
Four recommendations for future market fuel study practice represent a departure from prior practice:
- Fuels expected in the market should be used. For most efficient use of resources, effects resulting from a change of ethanol level should be evaluated by employing only fuels with the ethanol levels of interest, and with hydrocarbon (BOB) compositions that are expected in the market at those ethanol levels. Each ethanol level should be represented by one fuel, with average expected market hydrocarbon composition, or by a suite of fuels, with a distribution of hydrocarbon compositions that reflect expected market variability in composition. In this way measurement effort is not devoted to fuel formulations that may never enter commercial use, and nonlinear blending effects are addressed directly by the study fuels. This approach is dedicated to a singular objective, rather than the development of a broad model for fuel formulation impact. It also provides results that are a strong test for broad models.
- Fuel composition should be used to define fuels. For precise fuel description, measures of fuel composition are preferable to properties that are determined using standardized protocols. Composition varies widely, and properties alone are insufficient to define a fuel’s behavior. The fuel composition is unambiguous, whereas several fuels blended with diverse compositions may satisfy the same set of properties. Detailed hydrocarbon analysis (DHA) of fuels has become more reliable and rapid, permitting the grouping of molecular species by type, weight or expected influence. Use of composition to govern blending and fuel characterization avoids interference of non-linear blending effects on values of study parameters. Also, standardized protocols are not necessarily governed by the same chemistry, physics or time constants as the processes occurring in real-world injection, combustion, catalysis, adsorption and permeation. However, certain fuel properties, such as octane numbers, will continue to govern formulation of market fuels sold at the pump.
- Influence of vehicle technology merits higher recognition. Vehicles used in a study should be well characterized regarding their powertrain technology and control strategy, and not treated as “black boxes.” Although each automobile is a discrete species, the vehicle cohort may also be characterized by numerical descriptors, such as power to weight ratio, or engine power density, to appreciate the vehicle interaction with fuels at different loads and speeds. Prior studies have recognized major classifications, such as whether port fuel injection (PFI) or gasoline direct injection (GDI) is used, or if the engine is turbocharged. More thorough classification supports statistical analysis of vehicle-to-vehicle differences and defines the applicability of results from each vehicle for application to the on-road fleet. It is important that the practice of monitoring vehicle sensors and broadcast data during testing is adopted consistently and used in post-test evaluation of emissions effects. Understanding the role of engine control strategies in governing both emissions and fuel efficiency is of paramount importance. Involvement of automobile manufacturers or powertrain experts is beneficial in study planning and data interpretation.
- Driving schedules should mimic on-road use. During emissions testing, vehicles should be operated to mimic their real-world on-road use and environment as closely as possible if the intent is to predict emissions inventory. Unfortunately, the diversity of speeds, accelerations, climate and grades is excessive. Practically, it is recommended that data are gathered from vehicles includes operation at idle, light to medium load, and near full load, because modern engine controls will manage combustion differently in these circumstances. It is best to measure all emissions of interest continuously over a test schedule. Otherwise one should employ a low power cycle such as the FTP, a high power cycle such as the US06, and, if possible, a segment of wide open throttle (WOT) acceleration. Each of these loads may elicit performance, efficiency and emissions in different ways when challenged with different fuels.