This chapter will provide information on the development of the Greenhouse Gas Emissions Model (GEM) for vehicle certification of Class 4 through Class 8 vehicles, a description of the model, use of the model by vehicle original equipment manufacturers (OEMs), and analysis of the model with discussion of some of its limitations. Certification of Class 2b and Class 3 vehicles using a full chassis dynamometer is not addressed in this report, nor are changes to engine certification per Environmental Protection Agency (EPA) requirements for criteria pollutants, greenhouse gases (GHGs), and fuel use. Both subjects will be addressed in the final report.
From both technical and operational perspectives there is an objective to improve, and even optimize, efficiency metrics for trucks in each class. The existing National Highway Traffic Safety Administration (NHTSA) and EPA rules have already addressed the units to be employed for measuring engine efficiency, vehicle efficiency, and the associated carbon dioxide emissions levels, for both gasoline and diesel. From a practical standpoint, at the time of writing, the rules also drive change that provides economic benefit to the truck user. However, as truck efficiency standards advance, there are trade-off issues that must be addressed. The first is central to this report.
- Fuel efficiency versus greenhouse gases. If the change or optimization addresses reduction in fuel use, it may not necessarily address reduction in GHGs when a spectrum of fuels is considered. Different fuels have different carbon content, and they may also be associated with production of other GHG species. If fuel-insensitive efficiency and GHG rules are promulgated separately, two rules could drive toward two different fuel or technology options. For example, present-day natural gas engines may benefit GHG reduction more than diesel engines, but the inverse is true when it comes to efficiency. Dual-fuel engines increase the complexity because the balance of fuel use may be activity dependent.
There are also trade-offs with other national interests:
- Fuel efficiency or greenhouse gases versus cost. Reducing fuel use or GHG emissions may not be the most economically attractive scenario. Technology costs in some cases may exceed fuel savings over the vehicle life, and the least expensive fuel and technology combination may not offer the best efficiency or lowest GHG scenario. At a higher level, fuel choices may have substantial economic impacts beyond the trucking industry. For example, an advanced aerodynamic device that offers drag reduction of less than 1 percent is unlikely to offer payback during the first period of ownership if the weight and cost cross a certain threshold.
- Fuel efficiency or greenhouse gases versus criteria pollutants. Although reducing engine power output tends to reduce both fuel mass rate and criteria pollutant mass rate, some technologies designed to reduce criteria pollutants have adverse effects on engine efficiency and GHG production. In the technical realm, there is often a trade-off between these metrics. For example, a diesel particulate filter will reduce particulate mass but will raise fuel consumption due to exhaust back pressure and fuel used for necessary regeneration.
- Energy security versus efficiency and emissions. The use of alternative energy resources or a balancing of source uses may not yield highest efficiency, lowest GHG, or lowest criteria pollutants, but it may satisfy compelling national needs. For example, natural gas, as a domestic fuel, displaces imported oil. However, a spark-ignited natural gas engine is generally less energy efficient than a diesel engine.
- Technology impact. Scenarios may be made more complex when one fuel can be used in engines or power trains that employ two fundamentally different technologies if one technology offers benefit for one
metric and the other offers benefit for another metric. For example, natural gas may be used in compression-ignited or spark-ignited modes.
These metrics all have different currencies, and it is impossible to establish exchange rates between them from purely technical arguments. The balancing of these metrics is an issue of policy.
Recommendation 3.1: NHTSA, in consultation with EPA, should consider carefully the impact on related metrics when attempting to optimize for a single metric, or should otherwise establish a clearly articulated objective that weights, or places limits upon, relevant metrics.
The predecessor to the present committee, the National Research Council (NRC) Committee to Assess Fuel Economy Technologies for Medium- and Heavy-Duty Vehicles, published its final report (the “Phase One Report”) in 2010 (NRC, 2010). Appendix G of that report summarizes information on vehicle simulation tool requirements for regulatory use. Key requirements were outlined (NRC, 2010, pp. 221-225); they included
- Maximum reusability,
- Maximum flexibility,
- Selectable complexity,
- Code neutrality,
- Graphical user interface
—Select architecture, model, and data,
—Check model compatibilities to avoid crash or erroneous results,
- Select simulation type, including component evaluation, vehicle fuel efficiency, or drive quality
- Generic processes,
- Linkage with other tools,
—User access control,
—Database search, and
- Selection of single versus multiple tools for regulation.
The final rule was based, in part, on Recommendation 8-4 of the NRC Phase One Report.
Recommendation 8-4. Simulation modeling should be used with component test data and additional tested inputs from power train tests, which could lower the cost and administrative burden yet achieve the needed accuracy of results. This is similar to the approach taken in Japan, but with the important clarification that the program would represent all of the parameters of the vehicle (power train, aerodynamics, and tires) and relate fuel consumption to the vehicle task. Further, the combined vehicle simulation/component testing approach should be supplemented with tests of complete vehicles for audit purposes. (NRC, 2010, p. 190)
GEM Version 1, the proposed simulation tool, was provided to the industry for peer review. Comments were received and discussed in “Peer Review of the Greenhouse Gas Emissions Model (GEM) and EPA’s Response to Comments” (EPA, 2011a), which was published in August 2011, just prior to publication of the Phase I Rule. In addition, EPA (2011b) published a document to accompany the rulemaking, Greenhouse Gas Emissions Standards and Fuel Efficiency Standards for Medium- and Heavy-Duty Engines and Vehicles: EPA Response to Comments Document for Joint Rulemaking. This document included comments made by the OEMs and others on GEM and EPA’s responses. Some corrections to GEM were completed, and GEM Version 2.0 was released.
On September 15, 2011, EPA and NHTSA published their Final Rule on Greenhouse Gas Emissions Standards and Fuel Efficiency Standards for Medium- and Heavy-Duty Engines and Vehicles. GEM is a key part of the final rule. According to the Federal Register of that date,
compliance with the vehicle standard will typically be determined based on a customized vehicle simulation model, called the Greenhouse gas Emissions Model (GEM), which is consistent with the NAS Report recommendations to require compliance testing for combination tractors using vehicle simulation rather than chassis dynamometer testing. It is an accurate and cost-effective alternative to measuring emissions [of carbon dioxide] and fuel consumption while operating the vehicle on a chassis dynamometer as an indirect way to evaluate real-world operation and performance, various characteristics of the vehicle are measured and these measurements are used as inputs to the model. These characteristics relate to key technologies appropriate for this subcategory of truck—including aerodynamic features, weight reductions, tire rolling resistance, the presence of idle-reducing technology, and vehicle speed limiters. The model also assumes the use of a representative typical engine, rather than a vehicle-specific engine, because engines are regulated separately. Using these inputs, the model will be used to quantify the overall performance of the vehicle in terms of CO2 emissions and fuel consumption. The model’s development and design, as well as the sources for inputs, are discussed in detail in Section II below and in Chapter 4 of the RIA [Regulatory Impact Analysis]. (57 Federal Register 57116)
GEM is employed for model year (MY) 2014 vehicles and later. Given the rules for determining the model year of a vehicle, a 2014 model vehicle could be produced as early as January 2, 2013, or as late as December 31, 2014 (see Box 3-1).
EPA’s Regulations for Determining Model Year Designation
§ 85.2305 Subpart X—Determination of Model Year for Motor Vehicles and Engines Used in Motor Vehicles Under Section 177 and Part A of Title II of the Clean Air Act.
§ 85.2301 Applicability. The definitions provided by this subpart are effective February 23, 1995 and apply to all light-duty motor vehicles and trucks, heavy-duty motor vehicles and heavy-duty engines used in motor vehicles, and on-highway motorcycles as such vehicles and engines are regulated under section 177 and Title II part A of the Clean Air Act.
§ 85.2302 Definition of model year. Model year means the manufacturer’s annual production period (as determined under § 85.2304) which includes January 1 of such calendar year, provided, that if the manufacturer has no annual production period, the term “model year” shall mean the calendar year.
§ 85.2303 Duration of model year. A specific model year must always include January 1 of the calendar year for which it is designated and may not include a January 1 of any other calendar year. Thus, the maximum duration of a model year is one calendar year plus 364 days.
§ 85.2304 Definition of production period. (a) The “annual production period” for all models within an engine family of light-duty motor vehicles, heavy-duty motor vehicles and engines, and on-highway motorcycles begins either: when any vehicle or engine within the engine family is first produced; or on January 2 of the calendar year preceding the year for which the model year is designated, whichever date is later. The annual production period ends either: When the last such vehicle or engine is produced; or on December 31 of the calendar year for which the model year is named, whichever date is sooner. (b) The date when a vehicle or engine is first produced is the “Job 1 date,” which is defined as that calendar date on which a manufacturer completes all manufacturing and assembling processes necessary to produce the first saleable unit of the designated model which is in all material respects the same as the vehicle or engine described in the manufacturer’s application for certification. The “Job 1 date” may be a date earlier in time than the date on which the certificate of conformity is issued.
SOURCE: 49 CFR 565.15 and 60 Fed. Reg. 4738; January 24, 1995, unless otherwise noted.
Validation of the GEM was reported on the Regulatory Impact Analysis (EPA and NHTSA, 2011) as follows:
- Verification against two chassis dynamometer full combination-vehicle tests at Southwest Research Institute (SwRI) with 2 percent accuracy against three test cycles.
- Verification to another vehicle simulation tool, GT-Drive,1 with 2 percent accuracy. Ten vehicles were run against three test cycles.
On June 17, 2013, NHTSA and EPA published technical amendments to the rule in the Federal Register (78 Fed. Reg. 36370). Minor changes to the procedures for rounding in the GEM were provided, especially to align the different requirements of EPA and NHTSA. This is intended to make certain that a vehicle ends up in one, and only one, vehicle family/subfamily in both EPA and NHTSA designations. Coincident with this, NHTSA and EPA released GEM version 2.0.1 to the industry. Additionally, changes were made to the multiplication factor for advanced technology credits. This did not affect the GEM calculations or procedures for the baseline vehicle. Another change relates to the use of automatic engine shutdown (AES) technologies. EPA assumes 1,800 hours per year of idling. The technical amendment provides a requirement to discount the effect of the AES if it does not prevent 1,800 hours per year of idling. An error in coast-down testing done in preparation for the final rule was introduced into the definition of the aerodynamic bins.2 This was corrected in the technical amendment. Since this is an input to the GEM, it has a minor effect on the use of GEM in determining final results.
On August 16, 2013, NHTSA and EPA withdrew several of the technical amendments published on June 17, 2013.
1 The Phase I Rule references “GT-Power,” perhaps spuriously (76 Fed. Reg. 57146).
2 Binning is a method employed in model development to represent the real-world characteristics of a vehicle in a stylized, discretized manner. It involves the creation of a predefined set of notional categories into which real-world vehicles are sorted for purposes of carrying out the simulation. For example, GEM utilizes five bins to represent the aerodynamic characteristics of various vehicle configurations (EPA and NHTSA, 2011b, p. 2-46).
However, none of the withdrawals pertained to the use of GEM or the new version of GEM with revised rounding.3
The cooperative agreement between NHTSA and the NRC, modified in September 2013, includes the following element of the amended statement of task (Appendix B) that is of relevance to this chapter:
The committee will analyze and provide options for improvements to the certification and compliance procedures for medium- and heavy-duty vehicles—including the use of representative test cycles and simulation using various models—such as might be implemented in revised fuel consumption regulations affecting MY 2019-2022.
The committee has talked with several but not all vehicle manufacturers that use GEM. Based on those conversations, on its review of GEM, and on the need to acknowledge efficient technology broadly in a simulation model, the committee makes several recommendations in the present report. Execution of GEM requires insertion of data on aerodynamic properties and tire rolling resistance. These data are obtained from measurements, which are also discussed below.
The GEM model was developed for NHTSA’s and EPA’s Phase I Rule on medium- and heavy-duty vehicles (MHDVs) as a simplified method for determining the effects of the vehicle (rather than the engine) on fuel economy and GHG emissions. There are separate regulations focused on a certification of the engine as meeting established criteria for carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O). As such, it builds upon the work already done by EPA and the established procedures in 40 CFR Part 1065 to certify engines in a test cell for the criteria pollutants carbon monoxide (CO), nitrogen oxides (NOx), hydrocarbons, and particulate matter (see Box 3-2). The executable GEM program is based on MATLAB/Simulink,4 a common language for modeling and simulation in engineering. The user has five main inputs to the model:
- Aerodynamic drag coefficient value by bin,
- Steer and drive tire rolling resistance by value,
- Vehicle speed limiter presence and value,
- Vehicle mass reduction (lb), and
- Idle reduction presence.
Additional opportunities are classified as Innovative Technologies, which require a separate procedure and submittal process for taking advantage of the credits allowed. Hybrid technology is included in the Advanced Technology and Innovative Technologies categories. Natural-gas-powered vehicles, which will replace a fraction of the diesel-powered vehicles and which will be important in the coming years, are also included in the Advanced Technology and Innovative Technologies categories. Chapter 4 of the Regulatory Impact Analysis (RIA) that accompanied the promulgation of the Phase I Rule (EPA and NHTSA, 2011) provides an extended explanation of the model.
The standards for criteria pollutants are found at Title 40 Code of Federal Chapter 1037.102 entitled “Exhaust emission standards for NOx, HC, PM, and CO.” These pollutants are sometimes described collectively as “criteria pollutants” because they are either criteria pollutants under the Clean Air Act or precursors to the criteria pollutant ozone. These pollutants are also sometimes described collectively as “non-greenhouse gas pollutants,” although they do not necessarily have negligible global warming potential. As described in § 1037.102, standards for these pollutants are provided in 40 CFR part 86.
A brief overview is obtained by looking at the input screen of the model, shown in Figure 3-1. The output of GEM is cycle-weighted g/ton-mi CO2 and gal/1,000 ton-mi. An important part of understanding the model consists in examining the underlying assumptions, including the following:
- Fixed engine fuel map for one fuel type, diesel;
- Fixed manual transmission (10-speed for Classes 7 and 8 and 6-speed for vocational);
- Fixed axle ratios;
- Fixed tire sizes and a rolling resistance that is invariant with respect to speed and torque, though user supplied;
- Fixed electrical load;
- Fixed mechanical accessory power;
- Three fixed cycles (California Air Resources Board [ARB] Transient, 55 mph Cruise, 65 mph Cruise);
- Level roads;
- 1,800 hours engine idle;
- A fixed payload, although augmented with a weight reduction option; and
- A value for the product of area and drag coefficient, which is a user-supplied constant.
Since GEM is a relatively simple model focused on aerodynamics, rolling resistance, speed, weight, and idle control, it is not capable of acknowledging efficiency or
3 This change was “to address rounding inconsistencies when converting CO2 values to equivalent fuel consumption values in the Greenhouse Gas Emissions Model (GEM) simulation tool” (76 Fed. Reg. 36377).
FIGURE 3-1 Graphical user interface for GEM. SOURCE: EPA.
GHG emissions changes associated with the integration of advanced power trains, alternative fuels, hybrid and electric vehicles, and optimal component management. While tire rolling resistance is a key input, the assumption in GEM is that the tires are properly inflated per the recommendation of the tire manufacturer. In the real world, tires are not always maintained at the proper pressure. Tire pressure monitoring, maintenance, and control can only be addressed as Incentivizing Technology. Idle reduction input to the model is limited to a yes/no type answer, whereas idle reduction strategies in actual use can be much more sophisticated, especially as the latter relates to safety of the driver in hot and cold weather.
Certain items other than the few inputs to the model are handled as Incentivizing Technologies. NHTSA describes these in a presentation, shown in Figure 3-2.
To the committee’s knowledge, there had been little use of the track for Advanced Technology credits as of October 2013. Discussions with OEMs—Ford, GM, Navistar, Daimler Trucks, PACCAR, and Cummins—suggest three reasons for this:
- Volumes of product with the potential for Incentivizing Technology are low at this time;
- 2014-2016 requirements can likely be met without the need for Incentivizing Technology, while the need in 2017 is still open; and
- The procedure for proving effectiveness of the Incentivizing Technology is burdensome.
For reporting purposes, manufacturers are obliged to execute GEM to cover sales of Class 7 and 8 on-road tractors. The process is similar for vocational vehicles, but aerodynamic drag is not considered for those vehicles. Exception is granted for a limited number of vocational tractors, which are treated as vocational vehicles, as detailed in the following:
Class 7 and Class 8 tractors certified or exempted as vocational tractors are limited in production to no more than 21,000 vehicles in any three consecutive model years. (78 Fed. Reg. 36403)
Vocational tractors generating credits can trade and transfer credits in the same averaging sets as tractors and vocational vehicles in the same weight class. (78 Fed. Reg. 36403)
Off-road operation. Heavy-duty vocational vehicles including vocational tractors meeting the off-road criteria in 49 CFR 523.2 are exempted from the requirements in this paragraph (b), but the engines in these vehicles must meet the requirements of paragraph (d) of this section. (76 Fed. Reg. 57499)
FIGURE 3-2 Inclusion of incentivizing technologies in the Phase I Rule. SOURCE: NHTSA.
GEM is used by the manufacturer to compute projected load-specific fuel consumption for nine regulatory subcategories (see Table 3-1).
The manufacturer must execute GEM for each truck sold and must comply with the standards using the averaging, banking, and trading tools offered in 40 CFR Part 1066 (of the Phase I Rule). If it is the case that the manufacturer offers sufficiently efficient vehicles and that the purchasers, for economic or external reasons, elect to purchase these sufficiently efficient vehicles, then the GEM computations may not be time-sensitive, because compliance is assured. However, GEM computations must be completed to facilitate annual federal reporting. If the manufacturer foresees that compliance may be assured only by incentivizing the purchase of more efficient vehicles over less efficient vehicles, the manufacturer may need to execute GEM for sales control purposes. This may be done with a sense of urgency, to maintain a record of running averaged efficiency values throughout the reporting year. Separate execution of GEM for each vehicle is impractical using manual entry, which was required by the locked version that is publicly available. OEMs have, therefore, developed automated techniques, including obtaining source code to allow better implementation within their order entry system.
Finding: For manufacturers with high sales volumes, the gathering of sales data and their efficient and rapid processing with GEM is of great importance to facilitate annual
TABLE 3-1 Fuel Consumption of Class 7 and 8 Vehicles by Regulatory Category and Subcategory (gallons per 1,000 miles)
|Regulatory Subcategory and Effective Date||Type of Truck||Regulatory Category|
|Day Cab||Sleeper Cab|
|Class 7||Class 8||Class 8|
|MY2017 and later||Low roof||10.2||7.8||6.5|
|MY 2013-2015||Low roof||10.5||8.0||6.7|
SOURCE: 76 Fed. Reg. 57106-57513.
federal reporting. See Section “User interface, order entry, and GEM utility,” below, for additional information.
There is limited experience with the GEM model as the regulation is not yet in effect and only some OEMs have complied with the Phase I Rule, which takes effect in 2014 (Volvo Trucks, 2012; Daimler Trucks, North America, 2012; Reiskin, 2013). The validation of the model was limited to two vehicle tests and several computer runs against another simulation tool. Vehicle OEMs experienced with GEM have noted some errors that must be corrected.
It must be clear in considering GEM-based regulation whether the regulation is addressing national energy security, climate change, or economic benefit. This is of importance when alternative fuels are considered, and when any of these factors may be stronger influences than the regulation or run counter to the regulation. It would be optimal if GEM and the attendant rulemaking were capable of standing alone in driving the most appropriate technologies to raise fuel efficiency and if it did not rely on economic reality to ensure that inappropriate technology was not forced.
Future versions of GEM may need to take account of NHTSA’s and the Federal Motor Carrier Safety Administration’s (FMCSA’s) plans to require speed limiters on all current and new vehicles. The exact value of the speed limit, and the techniques for dealing with smart speed limiters that provide additional speed for passing, are not yet known (Miller and Cama, 2013).
The comments provided on GEM in this section are in most cases inseparable from the regulation implied by GEM and the processes used to find constants for GEM. In what follows, some comments and findings may refer to the tool itself, others to the measurement, design, modeling, and regulation implied by the tool. With the changes being considered by EPA and the recommendations in this report, GEM has the potential to become recognized as the best model for predicting vehicle performance on routes.
In general, the GEM user interface meets neither Recommendation 8-4 of Assessment of Technologies and Approaches to Reducing the Fuel Consumption of Medium- and Heavy-Duty Vehicles (NRC, 2010) nor the criteria established in Appendix G to that report. (These are discussed in the Section “Development of the Greenhouse Gas Emissions Model.”) OEMs have requested the source code for the model in order to adapt it to their order entry system and deal with proving that the outputs of the model are consistent with the original version (EPA, 2011b, p. 7-5). The cost and time to implement the model into order entry systems are considerable.
Finding: The current GEM Version 2.0.1 is very much simplified, has a poor user input interface that is not compatible with the automated order entry systems of the vehicle OEMs, and has insufficient output information.
Recommendation 3.2: The GEM programmers should make every effort to configure GEM to be compatible with existing OEM order entry systems. For maximum effectiveness, a vehicle OEM should be able to easily run GEM on the initial specification of a product from a customer and on any changes that may ensue by customer choice, manufacturer choice, or supplier availability of parts.
Finding: A sufficiently accurate version of GEM could be used by OEMs and fleets for making significant trade-offs on technology purchases for vehicles and for benchmarking their operations.
Recommendation 3.3: GEM should be made to provide a more useful output that includes graphs and other presentation methods that will allow for greater insight into the actions that an OEM can take to improve GHG emissions and fuel efficiency. To this end, future versions of GEM must be sufficiently sophisticated to yield realistic predictions of truck efficiency and accurate predictions of efficiency changes in response to design variables for a variety of vehicle activities.
Weight and Rolling Resistance
The major GEM inputs are for rolling resistance of the tires and for drag coefficient of the vehicle. Generally, the truck manufacturer obtains the tire data from the tire manufacturer, but it must either conduct aerodynamic coast-down testing in accordance with 40 CFR Part 1066 or use an alternative acceptable method to obtain drag area.
The weight reduction input in GEM is limited to a fixed set of technologies and parts. This does not successfully represent the efforts of OEMs to search for weight savings in all new part designs and might discourage innovation. The goal is to minimize the weight of the vehicle and maximize the weight of the load to achieve the optimum freight efficiency as measured in ton-miles while still meeting all required road, load, bridge, and safety requirements. Weight reduction credit might better be focused on achieving a total vehicle weight less than a given number.
In contrast to the well-established process for determining engine efficiency and GHG emissions, the GEM inputs for tire rolling resistance and aerodynamic drag are obtained using processes that were defined more recently and for
which the variability is less well documented.5 Obtaining tire rolling resistance values was initially reported to be a problem for OEMs (as the committee learned through presentations by Ford,6 GM,7 and Navistar8). Calibration of equipment traceable to national standards for passenger car tires is in place in Europe and self-calibration of equipment is in place for the grading of tires. Calibration of truck tire characterization equipment needs to be implemented in U.S. regulations.
Finding: For tire rolling resistance, there needs to be high confidence in input values to GEM, because fuel efficiency improvement rests on incremental changes in technology performance. Accuracy, as well as agreement between repeat tests and different laboratories, needs to be sufficiently good to permit detecting changes in performance.
Recommendation 3.4: A mechanism needs to be implemented for obtaining accurate tire rolling-resistance factors, including equipment calibration, and for maintaining that information in a public database. This might be managed in the same way that tread wear, temperature, and traction data are displayed through the federal Uniform Tire Quality Grading system.
Much has been learned about measuring aerodynamic performance since the NRC Phase One Report was published and the Phase I Rule promulgated. There needs to be high confidence in input values to GEM, because fuel efficiency improvement rests on incremental changes in technology performance.
Recommendation 3.5: Regulators should refine and improve the methods for obtaining aerodynamic performance data that better reflect real-world experience, including yaw and varying speeds.
Hard-Coded Features of GEM
GEM specifically does not allow for synergy between components, the operation or control of components in a most efficient way, or the engendering of efficiency through operation of a smaller component at higher relative load. In this way some opportunities for reduction in fuel use are lost. GEM specifies the performance maps for major components such as the engine and transmission but does not credit the vehicle manufacturer with benefits of using a potentially superior engine or transmission. Several presenters have suggested that full vehicle simulation or simulation of a power pack that includes the aftertreatment system and the transmission should be used, rather than only the engine. These approaches are intended to show the benefit of optimizing system-level operation rather than component-level operation. The idea of power pack versus engine only can apply either to test cells or to simulation. In this interim report, the committee supports the extension of current certification to include a power pack in a next-generation GEM simulation by the OEM. As indicated in the introduction to this chapter, the committee will consider full-vehicle simulation and other approaches to engine certification in the final report.
GEM is, accordingly, too simple to have direct value in the manufacturer’s vehicle design process. Rather, it influences design by virtue of its required use in demonstrating and achieving compliance. At this time of writing, the economic motivator of high fuel cost has as much influence as GEM on design, but it is inappropriate to rely on an economic argument to keep GEM from driving possibly unproductive outcomes.
Finding: In having limited inputs, in treating those inputs in a simple fashion, and in not allowing for skillful design integration, GEM (both the model and the regulation implied in using the model) may encourage designs that are suboptimal.
GEM employs a limited set of cycles to challenge the simulated truck. These cycles do not include real-world road grade. Being speed and time based, these cycles also do not allow for the faster acceleration of more powerful trucks, or the longer times that might be taken by less powerful trucks to complete some real-world routes. This deficiency is not evident when a hard-coded engine map, rather than a real OEM engine simulation, is used in the model.
Finding: Critical issues in the choice of test cycles in GEM include road-with-grade and different speed and torque profiles to better relate the test or model to real-world experience. In addition, route concepts or distance-based target schedules might provide a superior alternative to speed and time cycles.
Recommendation 3.6: The choice of test cycles and routes or schedules used in GEM needs to be readdressed thoroughly to avoid creating designs that are optimized for the
5 EPA has indicated that it has a contract with SwRI for “conducting SAE J1321 (fuel consumption benefit) testing on the road using verified tires.” Sam Waltzer and Cheryl Bynum, EPA, “SmartWay Technology Program: Influencing Efficient Freight Movement into the Future,” Personal communication to Tom Cackette and Chuck Salter, NRC Committee on Technologies and Approaches to Reducing the Fuel Consumption of Medium- and Heavy-Duty Vehicles, Phase Two, July 24, 2013.
6 Ken McAlinden, Ford Motor Company, “Heavy Duty GHG from a Full-Line Manufacturer’s Perspective.” Presentation to the committee, June 20, 2013.
7 Mark A. Allen and Barbara Kiss, “General Motors comments: NAS Panel on Heavy-duty GHG/CAFE Discussion,” Presentation to the committee, July 31, 2013.
8 Greg Fadler, Navistar, Inc., “Navistar Fuel Economy and Emissions,” Presentation to the committee, March 20, 2013.
test rather than for achieving real-world performance in the design process.
If OEMs were allowed to substitute any models or real test data for parts of GEM at will, a continuum for regulation would evolve between present-day GEM certification and whole-vehicle testing. This could be addressed by causing GEM to offer current practice default models for components, and, if a manufacturer’s component can exceed the performance of that default model, the model for the actual component could be substituted. Full vehicle modeling might also be addressed in the final report. Models should be capable of simulating real-world component behavior accurately and should not be oversimplified.
Recommendation 3.7: NHTSA should investigate allowing the OEM to substitute OEM-specific models or code for the fixed models in the current GEM, including substituting a power pack (the engine, aftertreatment, transmission). These models, whether provided by OEMs or fixed in the code, should be configured to reflect real-world operation accurately.
Certain parameters whose values can vary significantly in actual practice are fixed in GEM. As noted, GEM specifies the performance maps for major components such as the engine and transmission and does not credit the manufacturer with benefits for using a potentially superior engine or transmission, singly or in combination, such as occurs with predictive cruise control. Engines are certified separately in a test cell and measured at several points over specified drive cycles. In contrast, GEM addresses a steady-state condition for the speed and load of the vehicle. Therefore, the results may not represent the best fuel consumption over drive cycles that are expected to include different road speeds, grades, loads, and yaw angles.
Finding: GEM output is unaffected by the actual use of a smaller or larger engine in a truck in the same subcategory, because the engine map used by GEM is predefined. For example, downsizing of the engine is a known approach for saving fuel that GEM does not properly address.
Recommendation 3.8: The regulators should assess whether a steady-state speed-torque map is sufficient for GEM accuracy in engine efficiency prediction.
Similarly, GEM, in having a fixed test weight, does not differentiate between lighter duty Class 8 tractors that may pull volume-limited loads and heavier duty tractors that typically pull combined gross vehicle weights approaching 80,000 lb. Although single-axle sleeper cabs may offer an opportunity to reduce fuel use for volume-limited loads, they are not a subcategory in GEM.
Finding: The use of measured values can allow optimization of the engine/transmission/driveline (such as a power pack) to show its positive impact on the environment.
Recommendation 3.9: NHTSA, in coordination with EPA, should investigate substituting measured values for the fixed vehicle weights in the current model.
Controls, Vehicle Integration, and the Role of Aftertreatment
GEM and the implementation of GEM must encourage and facilitate all or most technology avenues that offer significant reduction in fuel consumption, by allowing accurate, quantitative modeling that does not restrict any fruitful options. Fuel consumption reductions that involve vehicle integration, power train integration, control strategies, auxiliary engine loads, aftertreatment strategies, or integrated power packs must be encompassed.
Finding: There are many technology avenues for reduction of fuel consumption that are currently not captured by GEM.
Recommendation 3.10: NHTSA should consider carefully ways in which a revised GEM simulation and associated test procedures can reflect the benefit of integrating an engine, aftertreatment, and transmission with interactive controls. NHTSA should consider whether a power train, or a power pack consisting of engine, aftertreatment, and transmission, can be certified for fuel use and GHG emissions in the manner of an engine; if that path is adopted, the power:weight ratio of the vehicle must be considered equitably.
Real-world operations and routes need to be considered for use in GEM, and fleet data could be used to develop or validate the target activity in GEM. Another source of information is the Federal Highway Administration Freight Performance Measures Initiative. As a guiding principle, operators of vehicles need to move freight safely and efficiently or otherwise perform work with a vocational vehicle. If properly designed, regulations and their implementation in GEM would not compel equipment that is unsuited to real-world operation, nor would GEM deprive a purchaser of performance features that are truly necessary for service but that may be atypical of current GEM cycles. For example, engine downsizing or lighter, lower friction, less robust transmissions might yield fuel savings but might not be appropriate for certain duty cycles.
The current supplementary emissions test cycle based on 65 mph and a 38,000 lb load might be good for providing a
certification cycle that is simple, but it does not reflect the real-world operation of over-the-road tractors that, according to reports from fleets, average less than 50 mph when taking into account regional operation and traffic. GEM should be capable of dealing equitably with, for example, a truck that cannot follow a speed-time trace owing to insufficient power over a part of the cycle or a truck so powerful that the cycle does not approach full power use.
Finding: GEM does not satisfactorily reflect real-world operation for over-the-road tractors.
Recommendation 3.11: NHTSA and EPA should modify the Greenhouse Gas Emissions Model (GEM) to employ cycles or vehicle activities that cover as large a fraction of over-the-road tractor operation as possible without becoming overly cumbersome. GEM should employ a sufficient number of truck types or subcategories to facilitate sound and beneficial regulation.
Tractor and Trailer
The tractor and trailer are fundamentally inseparable in addressing aerodynamic drag and design, but should also consider that present-day tractor and trailer fleets may limit realization of the benefits of integrated tractor and trailer design. If trailers are to be included in a future rule, GEM may need to be modified in some way to account for the interaction between the tractor and trailer. While a rule can be set to deal with the trailer alone, or a standardized trailer can be employed for aerodynamic testing, there are significant aerodynamic issues associated with the tractor-trailer gap and the air flow from under the tractor to under the trailer. NHTSA and EPA must avoid separate optimization of components, as this might create system-level issues and may prove counterproductive to true optimization. However, if the default is that a standard trailer must be used, that trailer must reflect best practice technology for efficiency improvement.
Finding: Using GEM to address all tractors in combination with trailers would accrue two principal benefits. First, low-rolling-resistance tires and aerodynamic devices are likely to offer benefits with tankers or flatbed trailers. Second, the current bobtail testing (i.e., operating a tractor without a trailer attached) of certain tractors is unlikely to faithfully represent their performance in combination with trailers.
Recommendation 3.12: NHTSA should assess the benefit of using GEM to address all tractors in combination with trailers.
Inclusion in GEM of Engines Using Alternative Fuels
Since natural-gas-powered vehicles are expected to become significantly more prevalent in the coming years, regulatory changes will likely aim at modifying GEM to incorporate these vehicles rather than driving them toward a separate compliance path. It would be preferable to use a model for the actual engine in GEM. Otherwise, the current procedures for advanced technology credits may be replaced by allowing the inclusion of a natural gas engine model in addition to a diesel engine model. (Natural-gas-fueled vehicles and the possible regulatory approaches to them are discussed in Chapter 5.) In a similar fashion, if any other fuel or combination of fuels shows a likelihood of substantial market penetration by 2017, it should be considered and modeled in the power train.
Information on Fuel Consumption
Other methods of obtaining fuel economy information could be investigated, such as making use of the following: any onboard computers required in the future by FMCSA; reliable reporting information already available to the government; and industry statistics that are regularly gathered, such as ton-miles and fuel taxes. Cooperation of major fleets in assessing real-world fuel economy would provide valuable validation or correction for GEM and raise public confidence in the fuel efficiency regulations. SmartWay fleets are required to provide data on an annual basis that could be of benefit.
On a monthly basis, each state is required to report to the Federal Highway Administration (FHWA) the number of gallons taxed by that state. These data are analyzed and compiled by FHWA staff. The data on the amount of on-highway fuel use for each state is then used to apportion federal revenue to each state. Yearly, the FHWA’s Office of Policy provides the previous year’s data for use in the attribution process. This allows the states extra time to review the data and verify that it is correct and ready to be used in attribution.9
Daimler Trucks. 2012. “North America: Daimler Trucks North America leads industry by certifying complete vehicle lineup GHG14 compliant,” http://media.daimler.com/dcmedia/0-921-657777-1-1466206-1-0-1-00-0-11701-1549054-0-1-0-0-0-0-0.html. Accessed November 15, 2013.
Environmental Protection Agency (EPA). 2011a. Peer Review of the Greenhouse Gas Emissions Model (GEM) and EPA’s Response to Comments. EPA-420-R-11-007. Washington, D.C.: EPA. August.
EPA. 2011b. Greenhouse Gas Emissions Standards and Fuel Efficiency Standards for Medium- and Heavy-Duty Engines and Vehicles: EPA Response to Comments Document for Joint Rulemaking.
EPA and National Highway Traffic Administration (NHTSA). 2011. Final Rulemaking to Establish Greenhouse Gas Emissions Standards and Fuel Efficiency Standards for Medium- and Heavy-Duty Engines and Vehicles Regulatory Impact Analysis. EPA-420-R-11-901. Washington, D.C.: EPA. August.
9 See http://www.fhwa.dot.gov/policyinformation/motorfuelhwy_trustfund.cfm. Accessed November 15, 2013.
Miller, E., and T. Cama. 2013. DOT’s speed-limiter proposal to add older trucks, aide says. Transport Topics. November 4.
National Research Council (NRC). 2010. Technologies and Approaches to Reducing the Fuel Consumption of Medium- and Heavy-Duty Vehicles. Washington, D.C.: The National Academies Press.
Reiskin, J.S. 2013. Navistar, Paccar continuing efforts to achieve 2014 GHG certification. Transport Topics. November 11.
Volvo Trucks. 2012. “Volvo trucks earn 2014 greenhouse gas certification for entire Class 8 vehicle lineup.” www.volvotrucks.com/trucks/na/en-us/_layouts/CWP.Internet.VolvoCom/Newsitem.aspx?News.Itemid=135785&News.Language=en=gb. Accessed November 15, 2013.