K
Model Description and Results for the EEA-ICF Model

METHODOLOGY OVERVIEW

The lumped parameter approach to fuel consumption modeling uses the same basic principles as all simulation models, but instead of calculating fuel consumption second by second, as is sometimes done, it uses an average cycle. Such an approach has been used widely by industry and regulatory agencies, most recently by the U.S. Environ mental Protection Agency (EPA) to help assess the 2012-2016 proposed fuel economy standards (EPA, 2008). The method can be generally described as a first-principles-based energy balance, which accounts for all the different categories of energy loss, including the following:

  • Losses based on the second law of thermodynamics,

  • Heat loss from the combusted gases to the exhaust and coolant,

  • Pumping loss,

  • Mechanical friction loss,

  • Transmission losses,

  • Accessory loads,

  • Vehicle road load tire and aerodynamic drag losses, and

  • Vehicle inertial energy lost to the brakes.

Conceptually, each technology improvement is characterized by the percent change to each of the loss categories. If multiple technologies are employed to reduce the same category of loss, each successive technology has a smaller impact as the category of loss becomes closer to zero.

EEA-ICF Inc.1 has developed a lumped parameter model that is broadly similar in scope and content to the EPA model (Duleep, 2007). In this model, all of the baseline vehicle energy losses are determined computationally, and many of the technology effects on each source of loss have been determined from data presented at technical conferences. However, the EPA does not document how the various losses were determined for the baseline vehicle: It says only that the vehicle has a fixed percentage of fuel lost to each category. The EPA also does not document how the technology-specific improvements in each category of loss were characterized. It appears that the losses for both the baseline vehicle and the effects of technology improvements were based not on computed values but on expert opinion.

MODEL COMPUTATIONS

Here the committee summarizes the EEA-ICF model. GM researchers Sovran and Bohn (1981) used numerical integration over the Federal Test Procedure city and highway driving cycles to determine the energy required at the wheel to move a vehicle over the driving cycle as a function of its weight, frontal area, drag coefficient, and tire rolling resistance coefficient. This procedure is used to compute the energy requirement at the wheel for the given baseline vehicle and translated to energy at the engine output shaft by using transmission and driveline efficiency factors (which differ by transmission type and number of gears) derived from the open literature. Accessory energy requirements are added as a fixed energy amount that is a function of engine size. This determines total engine output energy; average cycle power is then computed by distributing the energy over the cycle time when positive engine output is required—that is, the time spent at closed throttle braking and idle are accounted for separately. Average cycle RPM excluding idle was obtained for specific vehicles from simulation models on specific vehicles, and these data are scaled by the ratio of the N/V for the data vehicle and the baseline vehicle. The data are used to determine average brake mean effective pressure (BMEP) for the positive power portion of the cycle.

1

Energy and Environmental Analysis, Inc. (EEA) was acquired by ICF International during the course of this study. In this appendix, reference is made to EEA-ICF, although in the report as a whole reference is made simply to EEA.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 210
K Model Description and Results for the EEA-ICF Model METHODOLOGY OVERVIEW many of the technology effects on each source of loss have been determined from data presented at technical confer- The lumped parameter approach to fuel consumption ences. However, the EPA does not document how the vari- modeling uses the same basic principles as all simulation ous losses were determined for the baseline vehicle: It says models, but instead of calculating fuel consumption second only that the vehicle has a fixed percentage of fuel lost to by second, as is sometimes done, it uses an average cycle. each category. The EPA also does not document how the Such an approach has been used widely by industry and regu- technology-specific improvements in each category of loss latory agencies, most recently by the U.S. Environmental were characterized. It appears that the losses for both the Protection Agency (EPA) to help assess the 2012-2016 baseline vehicle and the effects of technology improve - proposed fuel economy standards (EPA, 2008). The method ments were based not on computed values but on expert can be generally described as a first-principles-based energy opinion. balance, which accounts for all the different categories of energy loss, including the following: MODEL COMPUTATIONS • Losses based on the second law of thermodynamics, Here the committee summarizes the EEA-ICF model. • Heat loss from the combusted gases to the exhaust and GM researchers Sovran and Bohn (1981) used numerical coolant, integration over the Federal Test Procedure city and high - • Pumping loss, way driving cycles to determine the energy required at the • Mechanical friction loss, wheel to move a vehicle over the driving cycle as a function • Transmission losses, of its weight, frontal area, drag coefficient, and tire rolling • Accessory loads, resistance coefficient. This procedure is used to compute • Vehicle road load tire and aerodynamic drag losses, the energy requirement at the wheel for the given baseline and vehicle and translated to energy at the engine output shaft • Vehicle inertial energy lost to the brakes. b y using transmission and driveline efficiency factors (which differ by transmission type and number of gears) Conceptually, each technology improvement is characterized derived from the open literature. Accessory energy require - by the percent change to each of the loss categories. If mul- ments are added as a fixed energy amount that is a function tiple technologies are employed to reduce the same category of engine size. This determines total engine output energy; of loss, each successive technology has a smaller impact as average cycle power is then computed by distributing the the category of loss becomes closer to zero. energy over the cycle time when positive engine output is EEA-ICF Inc.1 has developed a lumped parameter model required—that is, the time spent at closed throttle braking that is broadly similar in scope and content to the EPA and idle are accounted for separately. Average cycle RPM model (Duleep, 2007). In this model, all of the baseline excluding idle was obtained for specific vehicles from vehicle energy losses are determined computationally, and simulation models on specific vehicles, and these data are scaled by the ratio of the N/V for the data vehicle and the 1 Energy and Environmental Analysis, Inc. (EEA) was acquired by baseline vehicle. The data are used to determine average ICF International during the course of this study. In this appendix, refer- brake mean effective pressure (BMEP) for the positive ence is made to EEA-ICF, although in the report as a whole reference is power portion of the cycle. made simply to EEA. 210

OCR for page 210
211 APPENDIX K Fuel consumption is determined by the following mates, and the EPA concluded the results of their model relationship: were plausible, although a few technology packages required additional investigation. The EPA has indicated that it will IMEP = BMEP + FMEP + PMEP continue to use the lumped parameter approach as an analyti- cal tool, perhaps adjusting it to improve its fidelity as more where I is for indicated, F is for friction, P is for pumping, simulation results become available. and MEP is the mean effective pressure in each category. The EEA-ICF also performed analysis for the NRC Commit- fuel consumption model is derived from a methodology to tee on Assessment of Technologies for Improving Light- estimate an engine map using a semiempirical model devel- Duty Vehicle Fuel Economy (Duleep, 2008a, 2008b). Based oped by researchers at Ford and the University of Nottingham on the committee’s experience, when a number of engine, (Shayler et al., 1999). In this formulation, fuel consumption is transmission, and other technology improvements are simul- proportional to IMEP divided by indicated thermal efficiency taneously added to a baseline vehicle, the net fuel economy (sometimes called the Willans line), friction is determined benefit can be approximated by taking 90 percent of the addi- empirically from engine layout and is a function of RPM only, tive sum of the individual technology benefits, as developed and PMEP is simply intake manifold pressure (atmospheric by EEA-ICF. The committee used this technique to develop a pressure). Intake manifold pressure is solved for any given quick approximation of the level of agreement likely between BMEP, since IMEP is also proportional to intake pressure. the Ricardo simulations and the EEA-ICF lumped parameter This model explicitly derives thermal efficiency, friction loss, model. It was able to perform a quick analysis of only 23 of and pumping loss for the baseline vehicle. Fuel consumption 26 packages developed by Ricardo, since there were no data at idle and closed throttle braking are modeled as functions on HCCI engines, which were used in three of the Ricardo of engine displacement only. The baseline engine is always technology packages. modeled with fixed valve lift and timing, and the pumping Ricardo included one technology for which the committee loss is adjusted for the presence of variable valve timing if had no specific data. It called this “fast warm-up” technology applicable. The model can be construed as a two-point ap- because it involved the control of coolant flow to the engine proximation of a complete engine map and is a very reason- immediately after cold start. Based on the data presented able representation of fuel consumption at light and moderate by Ricardo, the benefit of the technology was estimated at loads where there is no fuel enrichment. 1 percent, including the benefit of the electric water pump. The technologies are characterized by their effect on each All other technology benefits were based on the data from of the losses explicitly accounted for in the model, and the ICF-EEA previous reports to DOE on fuel economy technol- representation is similar in concept to the representation in ogy. These benefit estimates were adjusted for the presence the EPA model. In the EEA-ICF analysis, the committee col- or absence of technologies on the baseline vehicle, since lected information on the effect of each engine technology on all benefits in the DOE reports have been typically defined peak engine efficiency, pumping loss, and friction loss as a relative to an engine with fixed valve timing and a four- cycle average from technical papers that describe measured speed automatic transmission. The results are illustrated in changes in these attributes from prototype or production Figure K.1, and the plot shows the difference between the systems. When these losses are not explicitly measured, they Ricardo results and the quick approximation method. are computed from other published values such as the change In 16 of the 23 cases, the Ricardo estimate is within +5 per- in compression ratio, the change in torque, or the measured cent of the quick estimate. In two cases, the Ricardo estimates change in fuel consumption. were more than 10 percent lower than the quick estimates, as shown in Figure K.1. In five cases, the Ricardo estimates were 10 percent (or more) higher than the quick estimate. The dif- Comparison of Results to Detailed Simulation Model ference implies that the benefits are larger than the simple sum Outputs of individual technology benefits and that technology syner- Both EEA-ICF and EPA have compared the lumped gies are positive. The committee also examined the technology parameter results with new full-scale simulation modeling packages in the two “low” and five “high” outliers. Both low results on several vehicle classes with different combinations outliers had technology packages with a continuously variable of planned technological improvements. The simulations transmission (CVT) as one of the technologies. The five high were done by the consulting firm Ricardo, Inc., and docu- outliers had no major technology improvement in common. mented in a separate report (Ricardo, 2008). The Ricardo More detailed analysis was also done with the EEA-ICF work modeled five baseline vehicles (standard car, large car, lumped parameter model. Constraints on resources and time small MPV, large MPV, and large truck) and 26 technology allowed the committee to analyze only 9 of the 23 cases combinations, covering gasoline and diesel power trains used with the lumped parameter model, but the 9 cases included in the EPA model, but there was no simulation of hybrids. both high and low outliers from the previous analysis. Three In a majority of the comparisons done by EPA, the lumped technology packages were analyzed for a standard car, parameter model estimates were close to the Ricardo esti- which used a Toyota Camry baseline; three for a compact

OCR for page 210
212 ASSESSMENT OF FUEL ECONOMY TECHNOLOGIES FOR LIGHT-DUTY VEHICLES 20 15 10 RICARDO-EEA % 5 0 –5 10 20 30 40 50 –10 –15 –20 EEA QUICK ESTIMATE % FIGURE K.1 Comparison of the difference between the Ricardo, Inc., results and the quick approximation method. Figure K-1.eps Comparison of Model Results to NRC Estimates van, which used a Chrysler Voyager baseline; and three for a standard pickup, which used a Ford F-150 baseline. The NRC study has developed a series of technology Table K.1 shows the results and compares them with those paths whose combined effect on fuel consumption was es- of the quick method. The more detailed modeling reduced timated from expert inputs on the marginal benefits of each the average difference between the Ricardo estimates and the successive technology given technologies already adopted. committee estimates for the Toyota Camry and the Chrysler Paths were specified for five different vehicles: small cars, compact van but increased the difference for the Ford F-150 intermediate/large cars, high-performance sedans, body-on- truck. The largest observed difference is for Package 10 on frame small trucks, and large trucks. There were no substan- the Ford, where the baseline 5.4-L V8 is replaced by a 3.6-L tial differences in the paths or the resulting fuel consumption V6 turbo GDI engine and the downsizing is consistent with estimates across the five vehicles: All estimated decreases the 33 percent reduction that was used. in fuel consumption were between 27 and 29 percent for TABLE K.1 Comparison of Fuel Economy Improvements (in Percent) from Ricardo, Inc., Modeling, EEA-ICF �uick Analysis, and the EEA-ICF Model Vehicle Technology Package Ricardo Estimate EEA �uick Result EEA Model Result Toyota Camry Z 33.0 23.7 32.6 1 13.0 23.7 23.1 2 22.0 22.4 21.9 RMS difference 8.15 5.85 Chrysler Voyager 4 26.0 30.9 29.9 6b 35.5 33.3 35.5 16 41.0 28.5 36.6 RMS difference 7.85 3.39 Ford F-150 9 32.0 30.0 28.3 10 42.0 28.2 26.4 16 23.0 21.3 23.4 RMS difference 8.12 9.25 NOTE: RMS, root mean square difference between the EEA-ICF estimate and the Ricardo estimate. The differences seem to be in the same range as the differences between the EPA estimates with their lumped parameter model and the Ricardo estimates. It is also important to note that the EPA model results are more consistent with the results of the EEA-ICF model. The “low” Ricardo result for Package 1 on the Camry is also significantly lower than the EPA estimate of 20.5 percent fuel economy benefit, which is closer to the EEA-ICF estimate of 23 percent than to the Ricardo 13 percent estimate. Similarly, the high Ricardo estimate for Package 10 on the Ford F-150 is also substantially higher than the EPA estimate of 30.5 percent fuel efficiency gain, which is, in turn, higher than the committee estimate of 26.4 percent but much lower than the Ricardo estimate of 42 percent.

OCR for page 210
213 APPENDIX K spark-ignition engines and 36 and 40 percent for diesel en- TABLE K.2 Comparison of Fuel Consumption Reductions gines. Since the “performance sedan” and intermediate sedan (in Percent) for NRC Estimates and the EEA-ICF Model specifications were not very different, only the small car, one Spark Ignition Path NAS EEA-ICF intermediate car, and two trucks were simulated. Simulation Small car 27 26.7 was done for the spark ignition engine and the diesel engine Intermediate/large car 29 27.3 paths, but not for the hybrid path. BOF small truck 27 27.3 Table K.2 lists the model results versus the committee BOF large truck 29 26.2 estimates for the eight cases (four for spark ignition and four for diesel). In general, the model forecasts are very close Diesel path Small car 37 35.7 to but typically slightly lower than the forecasts of experts, Intermediate/large car 37 36.2 although well within the range of uncertainty included in the BOF small truck 37 36.6 committee estimate. Only one vehicle, the full-size truck, BOF large truck 40 36.5 shows a larger difference on the diesel path. Historically, NOTE: BOF, body on frame. the committee’s method of forecasting the marginal benefit of technology along a specified path has been criticized as potentially leading to an overestimation of benefits for spark ignition engines since it could lead to infeasible solutions if REFERENCES total pumping loss reduction estimated exceeded the actual pumping loss. The simulation model output’s explicit track- Duleep, K.G. 2007. Overview of lumped parameter model. Presentation to ing of the losses addresses this issue directly to ensure that the National Research Council Committee for the Assessment of Tech- no basic scientific relationships are violated. nologies for Improving Light-Duty Vehicle Fuel Economy on October Fuel consumption is decreased by reducing the tractive 26, Washington, D.C. Duleep, K.G. 2008a. EEA-ICF Analysis of Ricardo simulation outputs. energy required to move the vehicle (by reducing weight, Presentation to the National Research Council Committee for the aerodynamic drag, or rolling resistance), reducing losses to Assessment of Technologies for Improving Light-Duty Vehicle Fuel the transmission and drive line, reducing accessory energy Economy on February 26, Washington, D.C. consumption, or reducing engine fuel consumption during Duleep, K.G. 2008b. EEA-ICF analysis update. Presentation to the National idle and closed throttle braking. Fuel consumption can also Research Council Committee for the Assessment of Technologies for Improving Light-Duty Vehicle Fuel Economy on April 1, Washington, be reduced by increasing engine efficiency over the cycle, D.C. which is accomplished by increasing peak efficiency or by EPA (U.S. Environmental Protection Agency). 2008a. EPA Staff Techni- reducing mechanical friction and pumping loss. Figures K.2 cal Report: Cost and Effectiveness Estimates of Technologies Used through K.5 show the technology path steps and track the to Reduce Light-Duty Vehicle Carbon Dioxide Emissions. EPA420- reductions from both approaches separately, with the reduc- R-08-008. Ann Arbor, Mich. Ricardo, Inc. 2008. A Study of the Potential Effectiveness of Carbon tion in energy required to drive through the test cycle shown Dioxide Reducing Vehicle Technologies. Report to the Environmental on top and the engine efficiency shown below. Peak engine Protection Agency. June 26. efficiency actually decreases slightly due to turbocharging Sovran, G., and M. Bohn, 1981. Formulae for the tractive energy require- and downsizing, but the cycle efficiency increases from ments of the vehicles driving the EPA schedules. SAE Paper 810184. about 24 to 29 percent owing to reduction in pumping and SAE International, Warrendale, Pa. Shayler, P., J. Chick, and D. Eade. 1999. A method of predicting brake friction loss (blue part of the bar). The general trends are very specific fuel consumption maps. SAE Paper 1999-01-0556. SAE Inter- similar across all four vehicle types, but the key feature is national, Warrendale, Pa. that pumping and friction loss are not reduced to physically impossible levels for the solution.

OCR for page 210
214 ASSESSMENT OF FUEL ECONOMY TECHNOLOGIES FOR LIGHT-DUTY VEHICLES 0.35 0.33 0.31 0.29 0.27 ACC DRIVETRAIN kWH/mile 0.25 TRANSMISSION TORQUE CONV. TRACTION 0.23 0.21 0.19 0.17 0.15 BASE COMP. 4-VALVE GDI/TURBO ENG FRIC. CVVL DCT6 WT. REDUC RRC REDUC. ACC Technology FRICTION PUMPING ENGINE EFFICIENCY Figure K-2 top 0.35 0.33 0.31 0.29 Efficiency Percent 0.27 0.25 0.23 0.21 0.19 0.17 0.15 BASE 4-VALVE GDI/ ENG CVVL DCT6 WT. RRC ACC COMP. TURBO FRIC. REDUC REDUC. Technology FIGURE K.2 Technology path steps and reduction in energy required to drive through the test cycle (top) and the engine efficiency ( bottom), body-on-frame small truck. Figure K-2.eps

OCR for page 210
215 APPENDIX K 0.25 ACC 0.24 DRIVETRAIN TRANSMISSION TORQUE CONV. 0.23 TRACTION 0.22 0.21 kWH/mile 0.2 0.19 0.18 0.17 0.16 0.15 BASE LESS VVL+DCP ENG GDI/ DCT WT. RRC ACC COMP. ICP FRIC. Turbo REDUC REDUC. Technology PUMPING FRICTION ENGINE 0.4 EFICIENCY 0.35 Efficiency Percent 0.3 0.25 0.2 0.15 BASE LESS VVL+DCP ENG GDI/ DCT WT. RRC ACC COMP. ICP FRIC. Turbo REDUC REDUC. Technology FIGURE K.3 Technology path steps and reduction in energy required to drive through the test cycle (top) and the engine efficiency ( bottom), midsize sedan. Figure K-3.eps

OCR for page 210
216 ASSESSMENT OF FUEL ECONOMY TECHNOLOGIES FOR LIGHT-DUTY VEHICLES ACC 0.22 DRIVETRAIN TRANSMISSION TORQUE CONV. 0.2 TRACTION 0.18 kWH/mile 0.16 0.14 0.12 0.1 BASE VVL+DCP GDI/ ENG VVLT DCT WT. RRC ACC COMP. Turbo FRIC. REDUC REDUC. Technology FRICTION PUMPING ENGINE 0.4 EFFICIENCY 0.35 0.3 Efficiency Percent 0.25 0.2 0.15 0.1 BASE VVL+DCP GDI/ ENG VVLT DCT WT. RRC ACC COMP. Turbo FRIC. REDUC REDUC. Technology FIGURE K.4 Technology path steps and reduction in energy required to drive through the test cycle (top) and the engine efficiency ( bottom), small car. Figure K-4.eps

OCR for page 210
217 APPENDIX K ACC DRIVETRAIN 0.45 TRANSMISSION TORQUE CONV. TRACTION 0.4 0.35 kWH/mile 0.3 0.25 0.2 0.15 BASE 4V GDI- ENG VVLT DCT6 wt. REDUC. ACC COMP. TURBO FRIC.+OIL REDUC Technology FRICTION PUMPING ENGINE 0.4 EFFICIENCY 0.35 Efficiency Percent 0.3 0.25 0.2 0.15 BASE 4V GDI- ENG VVLT DCT6 wt. REDUC. ACC FRIC.+OIL COMP. TURBO REDUC Technology FIGURE K.5 Technology path steps and reduction in energy required to drive through the test cycle (top) and the engine efficiency ( bottom), full-size truck. Figure K-5.eps

OCR for page 210