SAE TECHNICAL
`PAPER SERIES
`
`981902
`
`Critical Issues in Quantifying Hybrid Electric
`Vehicle Emissions and Fuel Consumption
`
`Feng An and Matthew Barth
`University of California
`
`400 Commonwealth Drive, Warrendale, PA 15096-0001 U.S.A.
`
`Tel: (724) 776-4841 Fax: (724) 776-5760
`
`Future Transportation Technology
`Conference & Exposition
`Costa Mesa, CA
`August 11-13, 1998
`
`1 of 21
`
`FORD 1438
`
`

`
`The appearance of this ISSN code at the bottom of this page indicates SAE’s consent that copies of the
`paper may be made for personal or internal use of specific clients. This consent is given on the condition,
`however, that the copier pay a $7.00 per article copy fee through the Copyright Clearance Center, Inc.
`Operations Center, 222 Rosewood Drive, Danvers, MA 01923 for copying beyond that permitted by Sec-
`tions 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying such as
`copying for general distribution, for advertising or promotional purposes, for creating new collective works,
`or for resale.
`
`SAE routinely stocks printed papers for a period of three years following date of publication. Direct your
`orders to SAE Customer Sales and Satisfaction Department.
`
`Quantity reprint rates can be obtained from the Customer Sales and Satisfaction Department.
`
`To request permission to reprint a technical paper or permission to use copyrighted SAE publications in
`other works, contact the SAE Publications Group.
`
`All SAE papers, standards, and selected
`books are abstracted and indexed in the
`Global Mobility Database
`
`No part of this publication may be reproduced in any form, in an electronic retrieval system or otherwise, without the prior written
`permission of the publisher.
`
`ISSN 0148-7191
`Copyright 1998 Society of Automotive Engineers, Inc.
`
`Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE. The author is solely
`responsible for the content of the paper. A process is available by which discussions will be printed with the paper if it is published in
`SAE Transactions. For permission to publish this paper in full or in part, contact the SAE Publications Group.
`
`Persons wishing to submit papers to be considered for presentation or publication through SAE should send the manuscript or a 300
`word abstract of a proposed manuscript to: Secretary, Engineering Meetings Board, SAE.
`
`Printed in USA
`
`2 of 21
`
`FORD 1438
`
`

`
`Critical Issues in Quantifying Hybrid Electric Vehicle
`Emissions and Fuel Consumption
`
`981902
`
`Feng An and Matthew Barth
`University of California
`
`Copyright © 1998 Society of Automotive Engineers, Inc.
`
`ABSTRACT
`
`Quantifying Hybrid Electric Vehicle (HEV) emissions and
`fuel consumption is a difficult problem for a number of dif-
`ferent reasons: 1) HEVs can be configured in significantly
`different ways (e.g., series or parallel); 2) the Auxiliary
`Power Unit (APU) can consist of a wide variety of
`engines, fuel types, and sizes; and 3) the APU can be
`operated very differently depending on the energy man-
`agement system strategy and the type of driving that is
`performed (e.g., city vs. highway driving).
`
`With the future increase of HEV penetration in the vehicle
`fleet, there is an important need for government agencies
`and manufacturers to determine HEV emissions and fuel
`consumption. In this paper, several critical issues associ-
`ated with HEV emissions and fuel consumption are iden-
`tified and analyzed, using a sophisticated set of HEV and
`emission simulation modeling tools. Two different types of
`APUs are modeled, one based on a conventional gaso-
`line Internal Combustion Engine (ICE), the other based
`on a small hydrogen-fueled ICE. Different energy man-
`agement strategies and HEV configurations are exam-
`ined,
`including a parallel
`range-extender charge-
`depleting HEV, a series thermostatic charge-sustaining
`hydrogen HEV truck, and a power-splitting charge-sus-
`taining HEV (modeled after the Toyota Prius). Results
`show that HEV emissions and energy consumption have
`a high degree of dependency on: 1) the energy manage-
`ment strategy employed; 2) the length of the drive cycle;
`3) overall driving range; and 4) the initial battery state-of-
`charge (SOC). The simulation results present: 1) equiva-
`lent fuel economy; 2) emissions per mile; 3) pure electric
`range; and 4) total driving range, for the different cases
`analyzed. The simulation modeling tools are extremely
`useful for comparing different HEV configurations and
`should play an important role in developing a robust HEV
`emissions and fuel consumption test procedure.
`
`INTRODUCTION
`
`All hybrid-electric vehicles include three key components
`- an on-board auxiliary power unit (APU), an energy stor-
`age device (e.g., battery), and a HEV control system
`
`which carries out a particular energy management strat-
`egy. These three components can be highly variable from
`vehicle to vehicle and can have a profound effect on an
`HEV’s energy and emission performance. In this paper,
`we attempt to identify some of the key variables that have
`a significant effect on an HEV’s energy and emissions.
`Various HEV configurations, APU types, and control
`strategies are simulated and analyzed as examples to
`illustrate the energy and emission impacts.
`
`Specifically, an electric vehicle and three different hybrid
`electric vehicles are simulated using five different kinds of
`driving patterns (embodied as driving cycles). These
`modeled vehicles include: 1) a GM EV1 with NiMH batter-
`ies; 2) a parallel-configured UC-Davis-type range-
`extender charge-depleting HEV; 3) a series-configured
`charge-sustaining hydrogen-fueled HEV truck; and 4) a
`power splitting continuous-variable-transmission (CVT)
`HEV similar to the Toyota Prius. Further, a conventional
`vehicle (a 96 Toyota Corolla) is also simulated to serve as
`a baseline. The five driving cycles applied to the simula-
`tion models include: 1) EPA’s Highway Cycle (HWY); 2)
`EPA’s City Cycle (LA4); 3) California Air Resources
`Board (CARB) Unified LA 92 Cycle (LA92); 4) the New
`York City Cycle (NYC); and 5) the US06 Cycle (US06).
`
`Previous research has shown that conventional ICE vehi-
`cle tailpipe emissions and fuel economy are extremely
`sensitive to different driving cycles [1-7]. For example, a
`Ford Taurus tested using the LA4 cycle achieves approxi-
`mately 20 miles per gallon (MPG). By contrast, when
`applying the NYC cycle [1, 2] (which represents driving in
`congested urban conditions), the fuel economy drops to
`approximately 10 MPG [1]. Tailpipe emissions of pollut-
`ants including carbon monoxide (CO), hydrocarbons
`(HC), and oxides of nitrogen (NOx) also change dramati-
`cally with different driving conditions. Recent studies indi-
`cate that CO and HC emissions under aggressive driving
`conditions can be several orders of magnitude greater
`than when tested under LA4 certified conditions [8-11].
`
`Recent electric vehicle (EV) studies and tests also indi-
`cate that their range and efficiency depend greatly on
`operating conditions [12-14]. The “real-world” driving
`range for an EV may be somewhat different than what is
`
`1
`
`3 of 21
`
`FORD 1438
`
`

`
`advertised (based on the standard testing procedure),
`because “real world” driving patterns may include higher
`or lower speeds, more aggressive acceleration rates, and
`more stop-and-go type of driving.
`
`The estimation of an HEV’s energy use and emissions is
`more difficult, since HEVs can consist of a wide spectrum
`of vehicle classes, e.g., from a “range extender” to a
`“power assist”-type of vehicle (see, e.g., [15]). The vari-
`ables associated with an HEV design can be arranged
`and sized to meet different design objectives [15, 16].
`This design variability creates additional challenges
`when evaluating an HEV’s performance.
`
`MODELING APU EMISSIONS
`
`strategies where the APU is operated at a fixed torque-
`rpm point (e.g., a “thermostatic” control strategy), tran-
`sient emissions do not occur and are thus not a major
`concern.
`
`Table 1.
`
`Hydrogen Engine Fuel Consumption Map
`(bsfc vs. rpm & torque)
`
`RPM\Torq.
`
`10
`
`20
`
`30
`
`40
`
`50
`
`60
`
`1000
`
`1500
`
`2000
`
`2500
`
`3000
`
`3500
`
`0.2880
`
`0.1730
`
`0.1620
`
`0.1600
`
`0.1600
`
`0.1600
`
`0.2880
`
`0.1770
`
`0.1590
`
`0.1510
`
`0.1590
`
`0.1730
`
`0.2880
`
`0.2140
`
`0.2000
`
`0.1877
`
`0.1850
`
`0.1990
`
`0.2880
`
`0.2520
`
`0.2180
`
`0.2000
`
`0.1987
`
`0.2185
`
`0.2880
`
`0.2675
`
`0.2285
`
`0.2090
`
`0.2185
`
`0.2800
`
`0.2880
`
`0.2800
`
`0.2600
`
`0.2800
`
`0.2800
`
`0.2800
`
`Table 2.
`
`Hydrogen Engine NOx Emission Map (grams/
`second vs. rpm & torque)
`
`In this research, two different APU modeling methodolo-
`gies are used. A hydrogen-fueled ICE is modeled using a
`steady-state emissions map approach. In contrast, a gas-
`oline-fueled APU is modeled based on a modal emis-
`sions model developed
`for a conventional vehicle.
`Advantages and disadvantages of these two approaches
`have been discussed in another paper [4]. Essentially,
`the second modeling approach is more comprehensive,
`but requires more detailed measurement data. The
`engine-map approach is easier to implement, but is less
`accurate. These methodologies are described further
`below.
`
`HYDROGEN-FUELED ICE APU EMISSIONS MODEL –
`One of the simplest approaches to modeling an ICE’s
`energy consumption and emissions is based on measur-
`ing fuel consumption and emissions at steady-state loads
`and engine speeds, and creating what are called “engine
`maps”. These engine maps are essentially look-up tables
`with fuel consumption and emissions indexed as a func-
`tion of engine speed (RPM) and engine torque demand.
`
`When modeling a vehicle’s fuel consumption and emis-
`sions, it is necessary to first convert second-by-second
`vehicle velocity (and acceleration) into power demand,
`which then must be translated into second-by-second
`engine speed and torque demand. The engine maps can
`then be applied, giving the energy consumption and
`emissions by interpolating over the maps. This modeling
`methodology works well over slowly changing velocities,
`however it potentially can miss transient emissions that
`may occur during transitions from different steady-state
`levels.
`
`For our example vehicle, a brake specific fuel consump-
`tion (bsfc) map and a NOx emission map for the hydro-
`gen-fueled APU were created using on steady-state
`engine emission tests. In Table 1, RPM is engine speed
`in revolution per minute, torque is in lb-ft, and bsfc is in lb/
`hp-hr. In Table 2, NOx is in grams/second.
`
`The biggest advantage of the engine map approach is its
`simplicity. The major disadvantage is that it is based on
`steady-state measurements, thus may not accurately
`represent fuel consumption and emissions associated
`with transient events. Fortunately for many HEV control
`
`2
`
`RPM\Torq.
`
`10
`
`20
`
`30
`
`40
`
`50
`
`60
`
`1000
`
`1500
`
`2000
`
`2500
`
`3000
`
`3500
`
`0.0003
`
`0.0003
`
`0.0003
`
`0.0004
`
`0.0014
`
`0.0068
`
`0.0003
`
`0.0003
`
`0.0003
`
`0.0004
`
`0.0014
`
`0.0068
`
`0.0003
`
`0.0003
`
`0.0005
`
`0.0009
`
`0.0017
`
`0.0061
`
`0.0005
`
`0.0004
`
`0.0006
`
`0.0010
`
`0.0037
`
`0.0086
`
`0.0006
`
`0.0007
`
`0.0008
`
`0.0018
`
`0.0086
`
`0.0130
`
`0.0007
`
`0.0009
`
`0.0015
`
`0.0028
`
`0.0130
`
`0.0173
`
`Based on Table 1, the operational point with the lowest
`bsfc value (0.151 lb/hp-hr) is at torque 40 lb-ft @ 1500
`rpm. The corresponding NOx emission rate is 0.0004 g/s.
`This point is often referred to as the “sweet spot”. It is
`important to note that this sweet spot doesn’t correspond
`to the lowest emission rate, which is 0.0003 g/s based on
`Table 2. Ideally, one would like to operate the engine at its
`sweet spot which corresponds to both the lowest fuel
`consumption value and emission rate. In this case, how-
`ever, it would mean operating the engine at 40*1500/
`5252 = 11.4 hp, which is only about 1/3 of its rated power,
`which is unacceptable. In terms of functionality, many
`control strategies want to operate near the designed
`maximum power point. For this modeled hydrogen-fueled
`ICE, this corresponds to 50 lb-ft@3000 rpm, which is
`equivalent to 50*3000/5252 = 28.6 hp. This point corre-
`sponds to a bsfc value of 0.2185 lb/hp-hr, and a NOx
`emission rate of 0.0086 g/s. A more detailed discussion
`of the thermostat control strategy can be found in a later
`section of this paper.
`
`GASOLINE-FUELED ICE APU EMISSIONS MODEL –
`An alternative to the engine-map approach is developing
`a true “modal” emission model. A modal emission model
`predicts emissions (and fuel consumption) based on
`vehicle operating mode, e.g., idle, steady-state cruise,
`various levels of acceleration/deceleration, etc. A major
`advantage of modal emissions modeling approach is that
`it handles transient events more accurately. This can be
`
`4 of 21
`
`FORD 1438
`
`

`
`important for some HEV configurations, such as a paral-
`lel-configured HEV which allows its APU to be engaged
`in load-following driving, where transient driving events
`have direct impact on APU operation. The disadvantage
`of the modal emissions model approach is that it requires
`much more detailed testing data.
`
`Using this modal emissions modeling approach, vehicle
`tailpipe emissions are modeled on a second-by-second
`basis as the product of three components: fuel rate (FR),
`engine-out emission indices (gemission/gfuel), and cata-
`lyst pass fraction (CPF) [6, 8, 17]:
`
`detailed discussion on this modal emission modeling
`approach is given in [18].
`
`One of the advantages of the modal emissions modeling
`approach for evaluating HEV ICEs is that we can appro-
`priately “downsize” established models that were devel-
`oped from extensive testing from a parallel research
`program. Several parameters from the established mod-
`els (such as engine displacement, emission indices, etc.)
`were reduced to give fuel and emission responses typical
`of a smaller sized engine that might be employed as and
`HEV’s APU.
`
`) • CPF
`
`(Eq. 1)
`
`gemission
`gfuel
`Here FR is fuel use rate in grams/s, engine-out emission
`index is grams of engine-out emissions per gram of fuel
`consumed, and CPF is the catalyst pass fraction, defined
`as the ratio of tailpipe to engine-out emissions.
`
`tailpipe
`emissions
`
`= FR • (
`
`The general structure of the modal emissions model is
`composed of six modules, as illustrated by the six rectan-
`gular boxes in Figure 1: 1) engine power demand; 2)
`engine speed; 3) air/fuel ratio; 4) fuel-rate; 5) engine-out
`emissions; and 6) catalyst pass fraction. The model as a
`whole requires two groups of inputs (rounded boxes in
`Figure 1): A) input operating variables, such as the sec-
`ond-by-second speed trace; and B) model parameters,
`such as vehicle mass and engine size. The output of the
`model is tailpipe emissions and fuel consumption.
`
`There are four operating conditions in the model (ovals in
`Figure 1): a) cold start; b) stoichiometric; c) enrichment;
`and d) lean burn. Hot-stabilized vehicle operation encom-
`passes conditions b) through d); the model determines
`which condition the vehicle is operating at a given
`moment by evaluating vehicle power demand. For exam-
`ple, when the vehicle power demand exceeds a power
`enrichment threshold, the operating condition switches
`from stoichiometric to enrichment conditions. The model
`does not inherently determine when a cold start occurs;
`rather, the user must specify any cold start conditions.
`The model does determine when the operating condition
`switches from cold start to stoichiometric, however.
`
`The vehicle power demand (1) is determined based on
`operating variables (A) and specific vehicle parameters
`(B). All other modules require the input of additional vehi-
`cle parameters determined based on dynamometer mea-
`surements, as well as
`the engine power demand
`calculated by the model.
`
`The air/fuel equivalence ratio (which is the ratio of stoichi-
`ometric air/fuel ratio, roughly 14.6 for gasoline, to the
`instantaneous air/fuel ratio), f , is approximated only as a
`function of power, and is modeled separately in each of
`the four operating conditions a) through d). The core of
`the model is the fuel rate calculation (4). It is a function of
`power demand (1), engine speed (2), and air/fuel ratio
`(3). Engine speed is determined based on vehicle veloc-
`ity, gear shift schedule and power demand. A more
`
`3
`
`(A) INPUT
`OPERATING
`VARIABLES
`
`(B) MODEL
`PARAMETERS
`
` (2)
`ENGINE SPEED
` (N)
`
`(1)
`POWER
`DEMAND
`
` (4)
`FUEL RATE
` (FR)
`
` (3)
`AIR/FUEL
`EQU. RATIO
`
` (F)
`
`b. Stoichiometric
`c. Enrichment
`d. Lean
`
`a. Cold start
`
`TAILPIPE
`EMISSIONS
`&
`FUEL USE
`
`(5)
`ENGINE-
`OUT
`EMISSIONS
`
`(6)
`CATALYST
`PASS
`FRACTION
`
`Figure 1. Modal Emissions Model Architecture.
`
`ELECTRIC POWERTRAIN SIMULATION MODEL
`
`In order to estimate key characteristics of both EVs and
`HEVs, an electric powertrain simulation model has been
`developed (after [19]). The approach used relies on a few
`simplified analytic formulas and some key physical
`parameters. The flowchart of the electric powertrain is
`illustrated in Figure 2. There are four key components in
`an electric powertrain:
`
`1. a tractive power demand module which converts sec-
`ond-by-second vehicle velocity to power demand
`required at the wheels (or tractive power Ptractive);
`2. a drivetrain module which converts second-by-sec-
`ond tractive power Ptractive into second-by-second
`required motor torque (Tormotor) and required motor
`speed (rpmmotor);
`3. a motor/controller module which converts motor
`torque and speed into power required from the bat-
`tery terminal (Pbatt);
`4. a battery simulation module which calculates sec-
`ond-by-second current, voltage, and battery state-of-
`charge (SOC). Some of the key characteristics that
`are calculated include total energy used per distance
`(i.e., kWh/mi), electric range of the vehicle given bat-
`tery capacity, equivalent fuel economy, which also
`factors in power plant efficiency (i.e., MPGequiv), and
`overall battery efficiency.
`
`The input requirements for this simulation model (shown
`as rounded boxes in Figure 2) are categorized into four
`categories: 1) the second-by-second input operating vari-
`
`5 of 21
`
`FORD 1438
`
`MODEL EMISSIONS MODEL ARCHITECTURE
`

`
`ables (i.e., vehicle velocity and grade profiles); 2) vehicle
`parameters such as vehicle weight, aerodynamic drag,
`tire size, etc.; 3) a motor/controller loss map; and 4) bat-
`tery parameters.
`
`Figure 2. Electric Powertrain Simulation Model Flow
`Chart
`
`HEV SIMULATION MODELS
`
`The electric powertrain simulation model described
`above can be used for evaluating the operation of a pure
`electric vehicle. However, in order to evaluate hybrid elec-
`tric vehicles, it is necessary to integrate the APU opera-
`tion with associated fuel consumption and emissions.
`
`In this paper, we consider three different HEV configura-
`tions: 1) a series HEV configuration (see Figure 3a); 2) a
`parallel HEV configuration (see Figure 3b); and 3) a
`power-split HEV configuration (see Figure 3c).
`
`SERIES HEV SIMULATION MODEL – The series HEV
`simulation model flow chart is shown in Figure 3a. The
`model is composed of the following modules:
`
`1. a power demand module which calculates the engine
`power demand (Pdemand) (this module is a combina-
`tion of the first two modules of the electric powertrain
`model);
`2. the motor/controller module which converts engine
`demand power into power required by the battery/
`APU combination (Pbatt/APU), and also determines
`the amount of regenerative braking power that can
`be supplied to the batteries (Pregen);
`3. a control strategy module which allocates power
`resources between the APU and the batteries;
`4. a battery simulation module which calculates sec-
`ond-by-second current, voltage, and battery SOC.
`Other summary variables include total energy used
`per distance (i.e., kWh/mi), and electric range;
`5. an APU modal emissions module which calculates
`the energy consumption and emissions of the ICE.
`Note that in the series configuration, power from the
`APU can be used to directly recharge the batteries,
`indicated by Pcharge.
`Similar to the electric powertrain simulation model, there
`are several input requirements, shown as rounded boxes
`in Figure 3a: 1) the second-by-second input operating
`variables (i.e., vehicle velocity and grade profiles); 2)
`vehicle parameters such as vehicle weight, aerodynamic
`drag, tire size, etc.; 3) a motor/controller loss map; and 4)
`
`4
`
`battery parameters. Please refer to [18] for further details
`on the different modules.
`
`parallel
`PARALLEL HEV SIMULATION MODEL – The
`HEV simulation model flow chart is shown in Figure 3b.
`Similar to that above, the model is composed of the fol-
`lowing modules:
`
`1. a power demand module which calculates the engine
`power demand (Pdemand);
`2. a control strategy module which allocates power
`resources between the electric motor/controller &
`batteries, and the ICE;
`converts
`3. the motor/controller module which
`requested power from the control strategy into power
`required by the battery (Pbatt), and also determines
`the amount of regenerative braking power that can
`be supplied to the batteries (Pregen);
`4. a battery simulation module which calculates sec-
`ond-by-second current, voltage, and battery SOC.
`Other summary variables include total energy used
`per distance (i.e., kWh/mi), and electric range;
`5. an ICE modal emissions module which calculates the
`energy consumption and emissions of the ICE.
`
`Please refer to [18] for further details on these modules.
`
`POWER-SPLIT HEV SIMULATION MODEL – The
`power-split HEV simulation model flow chart is shown in
`Figure 3c (modeled after the Toyota Prius). This configu-
`ration is neither a parallel nor a series configuration. It is
`closer to the parallel configuration, but differs in that a
`planetary gear system combined with a starter/generator
`can transfer power between the ICE and electric motor,
`both which are coupled to the driveshaft. In this configu-
`ration, the ICE provides the primary power, with a power-
`split device (planetary gear with starter/generator) send-
`ing power to both the driveshaft and the electric motor.
`
`The key element for modeling the power-split HEV con-
`figuration
`is
`to model
`the planetary-gear/generator
`device. For this, the relationship between the axle speed
`(Naxle) and vehicle speed (Vel) can be expressed as:
`
`=
`Radius)*3.1416*(20.447*Vel Naxle
`
`(Eq. 2)
`
`where Radius is the radius of the vehicle tire. The axle
`speed can be further converted into motor speed (Nmotor)
`as follows:
`
`=
`
`axlermotorN*G N
`
`(Eq. 3)
`
`where Gr is the reduction gear ratio. The relationship
`between generator speed (Ngen) and engine speed
`(Neng) can be expressed as follows:
`r
`+
`
`r
`
`(Eq. 4)
`
`40001000...)N*)(1 - N*-(G Nengaxlergen ££=
`
`engN
`
`6 of 21
`
`FORD 1438
`
`

`
`So far the missing link for all these relationships is deter-
`mining the engine speed Neng. Assuming that the engine
`is always operated near WOT throttle, Neng can be
`expressed as a function of power demand Peng:
`
`+
`
`A*24B B- N2engAC-
`
`=
`
`(Eq. 5)
`
`(a) Series Configuration HEV Simulation Model Flow
`
`(b) Parallel Configuration HEV Simulation Model Flow
`
`and:
`
`(Eq. 6b)
`
`(Eq. 6a)
`
`A = 0.25*60/(Nm - Idle) ª 0.005
`B = 1 - 0.25*Nm/(Nm-Idle) ª 0.667
`(Eq. 6c)
`C = -1/(Qm*0.0085*g)*Peng ª 1.743* Peng
`where Nm = 4000 rpm is the maximum engine speed,
`Idle = 1000 rpm is the minimum engine speed, Qm = 75
`lb.ft is the maximum engine torque, and g represents the
`distance of operational points from the WOT torque
`curve. If g = 1, the engine will always be operated at WOT.
`Here we assume g = 0.9, which means that the engine is
`always operated at 90% of its corresponding WOT
`torque. The reason for this is to reduce noise and vibra-
`tion associated with operating at exact WOT points.
`
`Table 3 lists some basic characteristics of the MY98 Toy-
`ota Prius engine, as well as a MY95 Toyota Tercel engine
`for comparison purposes. This table also shows that the
`Prius engine has very low minimum bsfc values (215 g/
`kWh for Prius vs. 235g/kWh for Tercel), and higher maxi-
`mum efficiency (38% vs. 35%).
`
`Table 3.
`
`Characteristics of Toyota Engines
`
`Specification/Engine
`
`Tercel
`
`Prius DOHC
`
`No. of Cylinder
`
`Displacement (liter)
`
`No. of Vales/Cylinder
`
`Compression Ratio
`
`Fuel System
`
`4
`
`1.5
`
`2
`
`PFI
`
`4
`
`1.5
`
`4
`
`13.5:1
`
`EFI
`
`Max Power (kW @ RPM)
`
`69.4 @ 5,400
`
`58 @ 4,000
`
`(c) Power-Split Configuration HEV Simulation Model Flow
`
`Max Torque (lb*ft @ RPM)
`
`100 @ 4,400
`
`75 @ 4,000
`
`Figure 3.
`
`
`
`where r
` is the planetary gear ratio defined as number of
`generator gear teeth over number of engine gear teeth.
`The engine speed Neng is confined between 1000 to
`4000 rpm.
`Equation (4) shows that the generator speed Ngen stands
`between the axle shaft speed Naxle and engine speed
`Neng, thereby providing additional freedom to regulate
`engine speed Neng. In this sense, it serves as a continu-
`ous variable transmission (CVT) system [20].
`
`Vehicle Test Weight (lbs.)
`
`2,625
`
`3,333
`
`EPA MPG
`
` - Manual (City/Hwy)
`
` - Automatic (City/Hwy)
`
`Modeled Engine Efficiency
`
`Minimum bsfc (g/kWh)
`
`Max engine efficiency
`
`35.4/46.8
`
`29.4/38.0
`
`235
`
`35%
`
`55/52
`
`215
`
`38%
`
`Table 4 presents the modeled engine bsfc map in the for-
`mat of lookup table for the Prius engine. Figure 4 shows
`the simulated Prius engine performance map.
`
`5
`
`7 of 21
`
`FORD 1438
`
`

`
`Table 4.
`
`Engine bsfc map for a MY98 1.5 Toyota Prius
`engine
`
`500
`
`829
`
`515
`
`410
`
`358
`
`945
`
`699
`
`450
`
`367
`
`325
`
`8.3
`
`16.7
`
`25.0
`
`33.3
`
`1391 1836 2282 2727 3173 3618 4064
`
`603
`
`402
`
`335
`
`301
`
`281
`
`542
`
`371
`
`314
`
`286
`
`269
`
`516
`
`358
`
`306
`
`279
`
`263
`
`515
`
`358
`
`305
`
`279
`
`263
`
`520
`
`360
`
`307
`
`280
`
`264
`
`527
`
`364
`
`309
`
`282
`
`266
`
`539
`
`369
`
`313
`
`285
`
`268
`
`41.7
`
`50.0
`
`58.3
`
`66.7
`
`75.0
`
`326
`
`305
`
`290
`
`279
`
`265
`
`300
`
`284
`
`272
`
`263
`
`256
`
`268
`
`258
`
`251
`
`245
`
`257
`
`249
`
`243
`
`238
`
`253
`
`245
`
`240
`
`235
`
`253
`
`245
`
`240
`
`235
`
`254
`
`246
`
`240
`
`236
`
`255
`
`247
`
`241
`
`237
`
`257
`
`249
`
`243
`
`238
`
`This map is generated based on limited knowledge of the
`Prius engine characteristics, and our analytical engine
`fuel consumption model. The following lists some key
`features about the Prius engine map:
`
`1. Its peak efficiency is approximately 38% (or 215 g/
`kWh bsfc), comparing to 35% (235 g/kWh bsfc) of
`the Tercel engine.
`2. It has no fuel enrichment operation. This is evi-
`denced by the engine performance contour map pre-
`sented in [21]. This engine performance map shows
`that there is no closed efficiency island in the map,
`and engine efficiency keeps increasing towards wide-
`open throttle operation. This can only be achieved by
`eliminated enrichment operation. The half-closed
`contour lines indicate that engine friction increases at
`both ends of the engine speed spectrum.
`3. The Tercel engine also has 1.5 liter displacement,
`and is a very efficient engine. It has efficiency islands
`centered around 2000 rpm in a relatively low engine
`torque range. In contrast, the Prius engine has its
`most efficient area close to its maximum torque and
`around 4000 rpm; this makes the Prius engine an
`ideal candidate for a hybrid application.
`
`HEV OPERATIONAL STRATEGIES
`
`In recent years many hybrid vehicles have been pro-
`posed, built and tested. As mentioned previously, large
`variations in possible HEV designs exist, as well as how
`they are operated. For example, some HEVs can operate
`most of the time like an electric vehicle and use the APU
`to remedy the range limitations of the battery pack that is
`charged while the vehicle is in storage. On the other
`hand, some HEV designs may never be plugged in;
`although they are refueled like a conventional ICE-vehi-
`cle, they use an HEV drivetrain as a means to achieve
`new degrees of optimization for high-energy efficiency
`and lower emissions.
`
`Figure 4. Prius Engine Performance Map, asterisks
`represent the engine operational points.
`
`A detailed discussion of different HEV types and design
`categories is beyond the scope of this paper (see, e.g.,
`[15, 22]). However, it is important to note three key issues
`that guide any categorization (after [22]):
`
`• Charge-Sustaining vs. Charge-Depleting—it
`is
`important to determine whether the HEV can operate
`indefinitely without discharging the battery. If the
`HEV can operate and keep its battery charge at a
`specified level, it is referred to as a “charge-sustain-
`ing” HEV. If the charge cannot be maintained during
`operation, it is referred to as a “charge-depleting”
`HEV. If an HEV is charge-depleting, then the fuel
`economy and emissions cannot be determined
`based on fuel energy alone.
`• Off-Board vs. On-Board Charging—off-board
`charging refers to the case where an HEV is charged
`from the external power grid while the vehicle is in
`storage; this is in contrast to HEVs that derive their
`electrical energy solely from on-board charging via
`their APU.
`• ZEV Operation Capability—if an HEV is capable of
`operating entirely in electric-mode only (i.e., the APU
`is not used), then it is referred to as ZEV-operation-
`capable. On the other hand, the electric motor may
`be sized too small for practical driving speeds to
`allow for ZEV operation. In this case, the HEV is
`ZEV-operation-incapable.
`
`6
`
`8 of 21
`
`FORD 1438
`
`

`
`9 of 21
`
`FORD 1438
`
`

`
`LOAD-FOLLOWING CHARGE-SUSTAINING CONTROL
`STRATEGY – A load-following control strategy is proba-
`bly the most widely used HEV strategy. Instead of operat-
`ing the APU at a single operating point, a load following
`strategy allows the APU to operate within a range speci-
`fied by the power demand, vehicle speed, and/or engine
`speed. A load-following strategy is particularly suitable for
`an APU that does not go enriched at wide-open-throttle
`(such as Toyota’s Prius 1.5 little engine). In this case, the
`engine is always more efficient running closer to the
`WOT region. This can be demonstrated by our fuel con-
`sumption model [1]:
`
`h
`
`h
`(cid:215)»))(( +
`(cid:215)(cid:215)
`thdTorqueENkTorque
`f
`
`
`(Eq. 7)
`
`where h
` is the overall ICE APU efficiency, N is engine
`speed, k(N) is the friction factor in kJ (rpm*engine dis-
`placement) and is function of engine speed N, f is the
`fuel/air equivalence ratio as functions of both N and
`Torque, and Ed is engine displacement. k(N) increases
`with N significantly at both low and high engine speeds.
`Equation (7) can be best demonstrated by Figure 7 as
`follows:
`
`1. When engine torque and power tends to zero, engine
`efficiency tends to zero as well;
`2. When engine torque tends to its maximum torque,
`engine efficiency tends to its maximum value;
`3. When engine power or engine speed tends to its
`maximum value, engine efficiency starts to decline
`again. This is particularly true for engines that go
`enriched under high power conditions.
`
`Figure 7. Load Following Control Strategy
`
`In Figure 7, the top dashed line is the maximum APU effi-
`ciency (around 32-35% for most gasoline ICEs). The sec-
`ond dashed curve represents the EV-mode efficiency.
`The reversed U-curve represents APU efficiency as a
`function of power. This figure illustrates that it is most effi-
`cient to operate the APU between the lower and higher
`motive powers.
`
`Figure 7 also illustrates that the EV-mode efficiency is the
`lower limit of the HEV operation efficiency. The EV-mode
`
`8
`
`efficiency is somewhat subjective. From a full-cycle point
`of view, it should include battery round-trip efficiency,
`motor/controller efficiency, as well as power-plant effi-
`ciency, transmission/distribution efficiency and charger
`efficiency from wall plug to battery (for off-board charging
`case).
`
`In summary, the efficiency limit for a HEV is determined
`by the following four factors:
`
`1. the peak efficiency of the APU;
`2. the efficiencies of electric components, such as bat-
`tery (round trip charging/discharging), motor, genera-
`tor, controller, etc.;
`3. the efficiencies of regenerative braking;
`4. the relative size of the “high efficiency island” of the
`ICE, or the lower and higher motive power levels in
`Figure 7.
`
`The operational strategy for the power-split HEV (i.e.,
`Toyota Prius) is a load-following strategy. The asterisks
`on the engine map (Figure 4) represent the engine oper-
`ational points in the load following mode. The APU is
`turned off and batteries are engaged when power
`demand is below a motive power level. Here we assume
`that the lower motive power level corresponds to the 30%
`efficiency point of the Prius engine, which is approxi-
`mately 6 kW of power demand. When the power demand
`is larger than the maximum power rating of the engine,
`the battery is engaged to provide additional power. It is
`assumed that the battery SOC is maintained between 40-
`80%.
`
`When the battery’s SOC is below 40%, the engine starts
`to charge the battery on-board. Since the generator’s
`maximum power is 15 kW, we assume that the engine
`will provide a maximum charging power of 15 kW to the
`batteries. On-board charging is also available through its
`regenerative braking mechanism.
`
`ENERGY & EMISSION PERFORMANCE
`ANALYSIS
`
`In this section, the energy and emission performance of a
`conventional ICE-powered vehicle, a pure EV, and the
`three different HEV configurations are compared. Specifi-
`cally, a conventional model year (MY) 1996 Toyo