`Characterizing Hybrid Electric Vehicles
`
`Michael Duoba
`Center for Transportation Research
`Argonne National Laborato y
`9700 South Cuss Ave., Bldg. 362
`Argonne, IL 60439
`
`presented at the
`7th CRC On Road Vehicle Emissions Workshop
`San Diego, CA
`April 9- 1 1, 1997
`
`ABSTRACT
`
`Many problems are associated with applying test methods, like the Federal Test Procedure (FTP),
`for HEVs. Although there has been considerable progress recently in the area of HEV test procedure
`development, many challenges are still unsolved. A major hurdle to overcoming the challenges of
`developing HEV test procedures is the lack of HEV designs available for vehicle testing. Argonne
`National Laboratory has tested hybrid electric vehicles (HEVs) built by about 50 colleges and
`universities from 1994 to 1997 in annual vehicle engineering competitions sponsored in part by the U.S.
`Department of Energy (DOE). From this experience, the Laboratory has gathered information about the
`basics of HEV testing and issues important to successful characterization of HEVs. A collaboration
`between ANL and the Society of Automotive Engineer's (SAE) HEV Test Procedure Task Force has
`helped guide the development of test protocols for their proposed procedures (draft SAE 5171 1) and test
`methods suited for DOE vehicle competitions. HEVs use an electrical energy storage device, which
`requires that HEV testing include more time and effort to deal with the effects of transient energy storage
`as the vehicle is operating in HEV mode. HEV operation with electric-only capability can be
`characterized by correcting the HEV mode data using results from electric-only operation. HEVs
`without electric-only capability require multiple tests conducted to form data correlations that enable the
`tester to find the result that corresponds to a zero net change in SOC. HEVs that operate with a net
`depletion of charge cannot be corrected for battery SOC and are characterized with emissions and fuel
`consumption results coupled with the electrical energy usage rate.
`
`The submitted manuscript has been authored
`by a contractor of the U.S. Government
`under contract No. W-31-109-ENG-38.
`Accordingly, the U.S. Government retains a
`nonexclusive. royalty-free license to publish or
`reproduce the published form of this
`contribution. or atlow others to do so, for US.
`Government purposes.
`
`BMW1030
`Page 1 of 15
`
`
`
`DISCLAIMER
`
`Portions of this domment may be illegible
`in electronic image products. _ Images are
`produced from the be: available original
`document.
`
`BMW1030
`
`Page 2 of 15
`
`BMW1030
`Page 2 of 15
`
`
`
`INTRODUCTION
`
`For decades, researchers have been studying the Hybrid Electric Vehicle (HEV) as a means for
`increased efficiency and lower emissions in passenger vehicles. The enabling technologies for electric
`propulsion (such as high-power electronics and energy storage) have been developed sufficiently to bring
`electric vehicles (EVs) to the market in the last couple of years. Many key EV technologies are
`applicable to HEVs, thus accelerating HEV development. Within the next few years, production HEVs
`will likely be sold in the United States and overseas.
`Any new vehicle technology must be evaluated by applying appropriate test procedures to
`accurately measure and quantify its fuel efficiency and emissions for certification purposes and for
`engineering evaluations and comparisons. The merits of new HEV technology must be fully understood
`to justify development and production.
`Whereas conventional vehicles and EVs draw upon only one source of energy, an HEV has two
`on-board energy sources from which motive power is provided. The format and structure of the original
`Federal Test Procedure (FTP) was designed as an attempt to characterize on-road vehicle operation. The
`assumptions associated with short-cuts used in the FTP, although effective for conventional vehicles, are
`not necessarily compatible with the complexity of HEV operation and do not allow HEVs to be
`accurately characterized.
`Standardize test protocols must be modified and reconfigured to accommodate HEV designs.
`Developing these new HEV test procedures is an underestimated problem that will have an enormous
`impact how we the engineering community and regulatory agencies assess these potentially prevalent
`vehicles of the future.
`
`BACKGROUND
`The oil shock of the 1970s spawned interest in HEV technology as a means to combat our nation’s
`oil dependency by building a higher mileage vehicle. In the 1980s interest in HEVs continued as a
`means to meet air pollution reduction goals, and a variety of HEVs were built and evaluated;l,2,3
`however, no domestic manufacturer showed interest in producing HEVs. In 1990, when the California
`Air Resources Board ( C A B ) adopted its Low Emissions Vehicle regulations4, they were interested in
`HEVs, but the limited availability of these vehicles halted the development of comprehensive test
`procedures. The test procedure C A B 5 adopted was more or less a standard FTP dynamometer test
`operated at worst-case conditions of the HEV, during which the engine is working its hardest.
`In 1992, GM presented a proposal to the Environmental Protection Agency (EPA) for test
`procedures specially suited for HEVs. GM observed that, “Neither the best-case nor the worst-case tests
`alone are sufficient. A fair characterization requires at least two extremes and a rational scheme for
`weighting them6” Also in 1992, a paper by INEL described a test procedure7 that recommends testing
`HEV operation until a full charge/discharge cycle is observed and terminating the test at the same battery
`state-of-charge (SOC) as the test started (more discussion about testing concepts will be given in the
`body of the paper). In also in 1992, the Society of Automotive Engineers (SAE) assembled the Hybrid-
`Electric Vehicle Test Procedure Task Force consisting of representatives from industry, the national
`laboratories, the U.S. EPA, and other interested parties to formulate a standard practice for testing HEVs.
`SAE’s test procedure is draft 5171 18 and has been a living document undergoing several significant
`revisions over the past few years. In 1993, the U.S. Department of Energy (DOE) signed five-year
`contracts with Ford and GM (Chrysler has since joined the DOE HEV program) to cost share the
`development of a production HEV for the mass market, thus underscoring a real need for a standardized
`test procedure.
`
`2
`
`BMW1030
`Page 3 of 15
`
`
`
`Although we have seen HEV development efforts grow over the years, few HEVs have been
`available for testing. Studying HEV test procedure using only one or two prototypes would overlook
`many HEVs that represent a considerably large variety of possible designs. C A W has been
`experimenting with a prototype HEV from Mitsubishi, but this experience is limited to only a particular
`type of HEV, which was reflected in C A B ’ S earlier procedures (however, in 1996, CARE3 staff
`informally expressed interest in using the SAE J 17 1 1 procedure when it is completed).
`Since 1988, DOE, through Argonne National Laboratory (ANL) has been partnering with the
`major domestic automobile manufacturers to showcase the engineering efforts of the best colleges and
`universities in North America through Advanced Vehicle Technology Competitions (AVTC). Since that
`time, there have been five competitionst in which over 50 HEVs have been tested and evaluated.
`Competition events covering design and performance characteristics have included dynamometer testing
`for emissions and fuel economy by using HEV test procedure concepts. Testing a wide selection of HEV
`designs has been an excellent opportunity to learn about and develop hybrid test procedures. The
`information in this paper is based primarily on information gathered during the competitions.
`
`HEV OPERATION
`
`HEVs posses very elaborate drivetrains and potentially operate in entirely new and complex ways.
`In fact, these new operational capabilities have often confused discussions regarding the application of
`HEVs to the national fleet because usually the definitions used and the assumptions made about hybrid
`vehicles are too loose. Some HEVs operate most of the time like an electric vehicle (zero emissions
`vehicle, or ZEV) and use the engine to remedy the range limitations of the battery pack that charges
`while the vehicle is in storage. Some HEV designs may never be plugged in; although they are refueled
`like a conventional vehicle, they use an HEV drivetrain as a means to achieve new degrees of
`optimization for high energy efficiency and low emissions. In spite of all these differences, what these
`vehicles do have in common is an energy storage device (EDS): either a battery, ultracapacitor, or
`flywheel, that can store and release energy throughout operation in the HEV mode. This technology
`presents new challenges in vehicle testing and characterization.
`
`Test #1
`
`Test #2
`
`Test #3
`
`Test #4
`
`Test #5
`
`Test #6
`
`Test #7
`
`I
`
`SOC
`
`ASOC= +4.5%
`33.4
`MPG=
`J/MPG=
`0.030
`
`~~~~~
`
`-13.5%
`65.3
`0.01 5
`
`+4.5%
`33.4
`0.030
`
`+12%
`27.7
`0.036
`
`-7.5%
`49.5
`0.020
`
`-9%
`52.7
`0.019
`
`+la%
`24.4
`0.041
`
`Figure 1: Test-to-Test Variation in HEV Operation
`
`I994 and 1995 HEV Challenge, I995 DOE Advanced Student Hybrid (DASH) Challenge. 1996 American Tour de Sol, 1996
`Futurecar Challenge.
`
`3
`
`BMW1030
`Page 4 of 15
`
`
`
`A series of test results from conventional vehicles would more or less show the same data with
`some expected scatter. In contrast, hybrid operation yields markedly different results from test to test;
`the results vary because of effects of transient energy storage plus, perhaps, a host of other possible
`parameters unique to HEV operation. Figure 1 illustrates this point.
`This simplified example shows the results of a series of HEV tests while the vehicle is in the HEV
`mode. The SOC in this example is changing constantly, but over time it remains within an operational
`window. Energy taken from the ESD for propulsion power during some tests supplement fuel energy
`usage, which results in high MPG. Because of this operation, each individual test can only capture a
`small segment of the entire vehicle operation that we are trying to characterize.
`
`Classzfiing HEY Designs
`A lengthy (but worthwhile) discussion of HEV types and design categories is beyond the scope of
`this paper (see Ref. 9), but the types of HEVs that affect testing will be explored here.
`Discussions of HEV designs usually begin with an explanation of the two fundamental design
`configurations: series and parallel. Each configuration may be more conducive to a particular operating
`scheme, but in reality, the configuration of the HEV (series or parallel) has no bearing on vehicle testing,
`outside of such issues as testing a 4-wheel-drive HEV on a 2-wheel-drive dynamometer. The vehicle
`configuration can be more or less “black-boxed,” and the focus of our interest in testing lies in two
`fundamental operational distinctions:
`
`1. Can the vehicle operate in its hybrid mode indefinitely without discharging the battery?
`
`2. Does the vehicle have the capability to operate in electric-only mode for a significant
`amount of time (distance)?
`
`Question ( 1) relates to “charge-sustaining’’ or “charge-depleting” operation as defined in draft
`versions of SAE 5171 1. If an HEV cannot maintain charge, then fuel economy and emissions cannot be
`defined in terms of fuel energy alone. Aside from the issue of on- and off-board charging, an off-board-
`charging HEV may still be capable of maintaining charge, but it uses off-board energy for ZEV
`operation. No matter how an HEV is evaluated, vehicles that use off-board electrical energy must be
`treated differently than vehicles that derive their electrical energy from on-board charging. This “apples-
`to-oranges’’ comparison is important, but beyond the scope of this paper.
`Question (2) address the problem with applying an emissions and fuel economy test to a vehicle
`that, during some of its operation, does not use fuel or emit pollution. If, for example, an HEV’s engine
`is invoked at the end of a cycle or not at all during a particular test, the resulting data may prove
`unrepresentative.
`A popular vision of HEVs is that they all have electric-only capability. Although it may be
`possible to use the electric motor by itself to drive the vehicle, the motor may be sized too small for
`practical driving speeds, or the vehicle control strategy may never employ electric-only operation. Some
`HEV designs always have the engine on throughout their operation. Moreover, an HEV operating with
`the engine on all the time does not necessarily designate the vehicle as having the ability to do all of its
`electrical charging on-board. We can conclude that these two design distinctions are independent, which
`means that we can express the possibilities in a 2-by-2 matrix, as shown in Figure 2.
`The matrix shows graphs in each category box that describe the operation of the possible vehicle
`designs expressed in SOC vs. distance plots. The graphs are useful in showing different operational
`modes and tracking the energy in and out of the battery. The shaded sections of the graph indicate
`engine operation, the unshaded regions show ZEV-mode operation. Again, the discussion of these HEV
`design types does not require information about drivetrain configuration (series or parallel).
`
`4
`
`BMW1030
`Page 5 of 15
`
`
`
`The charge-sustaining HEV design with electric-only capability (HEV designs #1, #2, and #3) is
`also called a “dual-mode” or “all purpose”7 HEV because it can operate like an EV or use the engine for
`continuous HEV-mode operation. Typically, the HEV will charge while the vehicle is stored, then run in
`ZEV mode until the batteries are depleted to a pre-determined set-point where the engine is invoked for
`indefinite vehicle range. This HEV type was considered by the California Air Resources Board to
`receive favorable credits in allowable fleet emissions averages because of the benefits of expected zero
`emissions use5.
`
`Electric-Only Capability
`
`Incapable of Electric-Only
`
`0
`0
`v )
`
`1 .
`
`0
`0
`(I)
`
`3.
`
`0
`0
`v)
`
`6 .
`
`0
`$
`
`7.
`
`Distance
`
`Distance
`
`Distance
`
`Distance
`
`Distance
`
`0
`0
`v)
`
`4.
`
`0
`0
`v)
`
`8.
`
`0 2
`
`9.
`
`Distance
`
`Distance
`
`Distance
`
`Distance
`
`Figure 2: Classifying HEV Types for Testing.
`
`= Engine on
`
`HEV designs #1 and #2 operate as a ZEV until the SOC level has fallen too low, at which time the
`engine is used to provide motive power and/or recharge energy. HEV design #2 invokes the engine after
`ZEV operation, but it appears to keep it on during HEV operation. The rising and falling SOC in #2
`represents HEV operation; if however, the engine were used for all the motive power, then the SOC
`would remain constant. HEV design #3 does not have as long a ZEV range (and may never charge off-
`board), but it does have the distinction of operating for significantly long distances with the engine off.
`
`5
`
`BMW1030
`Page 6 of 15
`
`
`
`The category of charge-depleting HEVs that have electric-only capability (HEV designs #6 and #7
`in the lower left of Figure 2) do not have the capability to be driven indefinitely over the test cycle
`without depleting the ESD and must receive off-board charging to continue running with full power
`capabilities. HEV design #6 starts the engine after the ZEV range is depleted, as in designs #1, and #2;
`however, the engine does not appear to be providing enough charge (either because it is sized too small,
`or the control of energy flow is not effective) to keep the SOC from falling. HEV design #7 is not
`necessarily a practical vehicle, but it demonstrates various possibilities in the design category. HEV #7
`cycles the engine on and off like HEV designs #1 and #2, but it would appear that the vehicle is capable
`of providing charge with the engine on, but the control strategy does not allow the engine to stay on long
`enough to keep the ESD charged.
`Charge-sustaining HEVs incapable of significant electric-only travel (HEV designs #4 and #5 in
`the upper right of Figure 2) operate more or less like a conventional vehicle. This HEV design is a
`candidate for achieving the Partnership for a New Generation of Vehicles (PNGV) goal of an 80-mpg
`five-passenger vehicle. The vehicle can be driven indefinitely with the combustible fuel without the
`need for off-board electrical energy. HEV design #4 uses the engine full-time, but the SOC will fluctuate
`according to the transient energy utilization of the ESD for optimum overall fuel efficiency and
`emissions characteristics. In HEV design #5, the engine does shut down for short distances, thus it
`actually has ZEV capability. However, the electric-only distances in #5 is not significant, and for our
`purposes, this operation is more like the ZEV-incapable designs, than the ZEV-capable designs.
`The SOC vs. distance graph for an electric-launch parallel HEV looks like that shown in HEV
`design #6. The operation is as follows: at a stop, the engine is off, and the vehicle is launched with the
`electric motor until a particular speed and or power level is reached; then, the engine starts and the
`vehicle operates in a hybrid mode. Again, because this vehicle travels insignificant distances in ZEV
`mode, it is classified in the Incapable of Electric-Only category.
`Charge-depleting HEVs that lack significant electric-only operation are similar to the previous
`type, except these vehicles most likely do charge off-board and suffer from the same limitation as do all
`charge-depleting E-IEVs - they have limited range under desired power levels. Both HEV design #8 and
`design #9 cannot maintain the SOC, but #8 always uses the engine, while #9 does shut down the engine
`periodically.
`
`Transitional Modes
`A complication of classifying an HEV is the possibility that a vehicle switches control strategies to
`a different operational mode. HEV designs #I, #2, and #7 start out in ZEV mode, which can be
`considered a transitional mode because eventually the control strategy switches to HEV mode when the
`SOC drops too low. However, the vehicle may again switch to another energy management strategy later
`while in the HEV mode. For example, a vehicle may turn the engine on but continue to deplete charge
`until the SOC becomes a very low, at which time the control strategy increases engine power levels to
`become charge-sustaining.
`Based upon the HEV prototypes to date, the SOC is typically a primary cause for a change in
`control strategy. An HEV may switch how the engine and motor work together based other inputs, like
`particular speeds or desired power levels, but for standardized testing, the response to these inputs will
`occur every time the test reaches that part of the speed trace. This is considered part of the same
`operational mode. If, however, an HEV reacts differently during a test to the same speed and or desired
`power level when started at the same initial SOC, this indicates a change in operational mode.
`Figure 3 shows four examples of vehicles that pass through transitional operational modes. HEV
`design # 10 acts like a charge-depleting HEV until the SOC drops and the HEV mode becomes charge-
`sustaining. This operation is a practical alternative to a charge-depleting design. For most commuting
`days, the HEV operates under an efficient and emissions-friendly charge-depleting control strategy;
`however, if the vehicle is called upon for extended trips, the vehicle is not range limited. In HEV design
`
`6
`
`BMW1030
`Page 7 of 15
`
`
`
`#12, the vehicle operates with a charge-depleting engine on/off strategy until the HEV mode switches to
`a strategy that keeps the engine on.
`HEV designs #11 and #13 continuously change their control strategy according to SOC. Like
`HEV designs # I O and #11, they charge deplete to a point, but the transition between charge-sustaining
`and charge-depleting is not as distinct.
`
`0
`0
`v)
`10.
`
`0
`
`8
`
`12.
`
`Distance
`
`Distance
`
`u
`0
`v)
`
`11.
`
`0
`
`s
`13.
`
`Distance
`
`Distance
`
`I
`
`Figure 3: Transitional Modes
`Other Mode Changes - There can be any number of parameters other than SOC that cause a change in
`control strategy. A vehicle with several control schemes that change will be very difficult to
`characterize. However, if the vehicle eventually reaches a more stable mode, then it can be characterized
`according to distance or time traveled in the transitional modes separately from the more predominant
`operational mode.
`
`I
`
`CHALLENGES IN HEV TESTING
`The discussion of HEV design types is useful for explaining why HEVs present such a challenge in
`testing. Fully characterizing such a complicated system requires more time and effort; to compound the
`problem, the wide variety of possible HEV designs cause unique challenges for each design type.
`
`SOC Measurement
`We begin the discussion of the difficulties in testing HEVs with the actual means by which the
`battery SOC is measured and tracked. Although we rely upon SOC to help characterize the vehicle’s
`operation, we can also have difficulties in the actual measurements that help track the EDS SOC
`accurately. The reason is that SOC cannot be measured directly; we can only track what passes in and
`out of the batteries, not what is currently inside. Battery testers describe the battery pack as a “rubber
`bucket,” - you never really know how much is actually in it.
`However, the parameter that does correlate most with SOC is ampere-hours, which is a measure of
`the amount of charge that has passed in and out of the ESD. The integrated amperes is essentially the
`quantity of charge - electrons going in and out of the battery. Tracking the energy in and out of the
`battery is not particularly useful because we are more interested in how much charge is left in the EDS;
`all energy going in and out of the pack is subject to irreversible losses because of the electrical resistance
`of the ESD. Although there is some hysteresis involved with tracking Ah because of various
`electrochemical mechanisms, the losses are smaller than those in tracking energy (kWh) and can be
`considered insignificant if the changes in SOC during testing are relatively small.
`
`7
`
`BMW1030
`Page 8 of 15
`
`
`
`Measuring the ampere-hours is a simple procedure. The main battery cables need a current-
`sensing device that can be logged and integrated with respect to time. The sample rate need not be very
`high; around 6 to 10 Hz will sufficiently account for the currents going in and out of an HEV ESD during
`testing.
`
`Transient Energy Storage
`Because HEV operation involves continuously storing and using energy from the ESD, we must
`consider this energy use in conjunction with the combustible fuel. Looking back at the example in
`Figure 1, the SOC is constantly changing, but over time, it remains within the operational limits directed
`by the vehicle control strategy, As we can see, each individual test captures only a small segment of the
`entire vehicle operation. The MPG measurement of any single test will not give us useful results if they
`are not coupled with the SOC and other test data. However, we could average many tests together to
`mitigate the transient effects of the ESD on the fuel economy results. For simulation studies, using
`numerous test runs is a useful approach, but for actual vehicle testing, this method may be impractical.
`An “SOC Correction” method was developed by the SAE Task Force to deal with this problem. This
`correction correlates the fuel usage data with the SOC data so that a result corresponding to a ASOC = 0
`can be found. Figure 4 graphically shows the ASOC = 0 result (using the data from Figure I) found at
`the intersection of the correlation line with the ASOC axis.
`
`0.045
`
`0.04
`
`0.035
`
`0.03
`
`0.025
`
`n 5 0.02
`
`0.015
`
`0.01
`
`0.005
`
`0
`-20
`
`-15
`
`-10
`
`,
`5
`
`I
`0
`5
`Delta SOC (v-)
`
`I O
`
`’ 15
`
`20
`
`Figure 4: SOC Correction Graph of Figure 3 Data
`
`The straight line in Figure 4 consists of fictitious data to illustrate this SOC correction concept. Actual
`test data will most likely show random scatter, with possibly some systematic curve shape or
`discontinuities that deviate from a linear correlation. Because comprehensive test data from real HEVs
`are rare, proving or disproving the effectiveness of this method is difficult, but small changes in SOC
`would reduce the possible error. A plot of many tests of a particular HEV would show the overall
`correlation and help locate the intersection point on the basis of a few skeleton points - but, again, too
`many required test runs would render such a procedure impractical.
`
`Engine On/Off Operation
`
`8
`
`BMW1030
`Page 9 of 15
`
`
`
`Perhaps the most problematic characteristic of some HEVs is engine onloff behavior found in
`some ZEV-capable HEV designs. The control strategy of HEVs manages the energy and power from the
`engine and electric motor for the most optimized performance, and this strategy may include periodic
`engine shutdown or delayed engine start-up. This new dimension in vehicle operation (compared with a
`conventional vehicle) is an important area of concern in applying standard test procedures to HEVs. As
`with other HEVs, added test redundancy can help characterize the operation; however, when the engine
`remains off for a significant amount of driving, profound anomalies may emerge. The anomalies are
`especially problematic emissions are analyzed.
`If we employ any kind of SOC correction strategy or test any less than a large, statistically
`significant string of tests, we run the risk of generating misrepresentative data caused by periodic or
`delayed engine starts. Consider the example in Figure 5.
`
`Test # I
`
`Test #2
`
`Test #3
`
`Test #4
`
`Test # I
`
`Test #2
`
`Test #3
`
`Test #4
`
`0
`0
`(I)
`
`a.
`
`Distance
`
`0
`
`b.
`
`Distance
`
`Figure 5: Engine OdOff Operation During HEV Testing
`
`The test procedure consists of four tests in series to determine the emissions and fuel economy.
`Although these two HEVs are nearly identical and would show very similar on-road emission rates, the
`test results do not reflect this fact. The differences in operation are that HEV(a) is very close to an
`engine-start at the end of the test sequence, while HEV(b) has initiated engine-start just before the end of
`the last test. SOC correction correlations indicate an engine charging rate that will account for the added
`charging that took place at the end of the test. However, for the emissions calculations of HEV(b), high
`start-up emissions may be associated with the small amount of charging that occurred at the end of the
`last test. For HEV(b), a slightly shorter test cycle will result in significantly different test results; thus,
`this HEV test procedure has not accurately characterized these vehicles as nearly identical.
`
`Incompatibilities with Standardized Test Procedures
`Hot- and Cold-Start Weighting - Engine odoff operation also causes problems with FTP hotkold-start
`weighting schemes. The FTP employs a weighting scheme for the cold- and hot-start tests to process the
`data to be representative of actual on-road usage. Driver-behavior statistics of an average number of
`daily trips, trip lengths, and soak times were analyzed, and a two-trip test was used for the FTP. The first
`trip includes the effects of a coId-start test (after an overnight rest period), and the second accounts for all
`starts after resting. The 10-min soak time between tests was based on the driver statistics of average
`number of daily trips and average soak times between trips. The emissions and fuel consumption data
`from the cold and hot tests are combined with a 43% / 57% weighting between the cold and hot start
`tests, respectively. Where E is the emissions rate or fuel consumption rate, the equation is as follows:
`
`Eq. 1
`
`An unmodified FTP cannot be used directly to test HEVs because of the configuration of the segments
`that make the FTP a three-bag test. The FTP test is actually based off two full Federal Urban Driving
`
`9
`
`BMW1030
`Page 10 of 15
`
`
`
`Schedules (FUDS) cycles. The last part of the second FUDS (after the first 505 s) is not tested because
`this phase (called “hot stabilized”) is considered redundant with the same portion of the first FUDS
`cycle. A decision was made to truncate the FTP by eliminating the redundant test segment (the last 866
`s of the second FUDS) to reduce the total test time. The entire bag weighting must be reworked to avoid
`significant test anomalies. The simple remedy for this problem is proposed in the early drafts of the SAE
`5171 1: run both FUDS cycles without the truncation.
`However, an engine on/off HEV will have problems with bag weighting as illustrated in Figure 6.
`Here, the FTP tests are modified to include the remainder of the last FUDS cycle, but notice that in test
`(a), the cold-start FUDS test does not include any emissions or fuel consumption data; thus, the hothold
`weighting would be undefined.
`
`I
`
`2
`
`s
`
`3
`
` 4
`
`1
`
`2
`
`s
`
`3
`
` 4
`
`1
`
`2
`
`s
`
`3
`
` 4
`
`r
`
`a.
`
`I
`
`Distance
`
`b.
`
`Distance
`
`0
`$
`
`C.
`
`Distance
`
`I
`
`Figure 6: Modified FTP to Have 4 Bags.
`
`In test (b), the engine shuts down quickly after start in the first test, then starts again in the second test,
`most likely resulting in high emissions from multiple start-ups. The results from this test may be
`unrepresentative, because longer test cycles would allow more warm engine operation, and shorter test
`cycles would exclude the engine start - both resulting in lower g/mi emissions rates. This phenomenon
`will be discussed in more detail in the following subsection.
`Trio Lenahs and Soak Times - The resulting emissions rates of any vehicle are sensitive to test length
`and soak time. Because a large proportion of the total tailpipe mass emissions from a test are expelled
`during warm-up, the actual test length will effect the g/mi emissions rate for the whole cycle (a longer
`test would result in a lower emissions rate). The soak time between tests (like the 10-min soak time
`given between tests in the FTP) also changes the final results by affecting the results of the subsequent
`engine start. A longer soak time will result in a higher mass emissions from the hot-start. These effects
`are fairly well known and predictable for conventional vehicles. The test length and soak time for
`standard testing (such as the FTP) is one scenario by which we gather information on all vehicles for
`emissions and fuel economy. Although this test is compromising, for a conventional vehicle, it is more
`or less representative of a broad range of possible test lengths and soak times. However, an HEV is not
`as easily characterized with such a specific and arbitrary test structure. An HEV with engine odoff
`operation is much less predictable than a conventional vehicle in its response to varying trip lengths and
`soak times. During actual vehicle operation of varying trip lengths and soak times, an HEV may have
`respectable emissions rates, however, if the test cycle forces one arbitrary and unrepresentative response,
`the test results may mistakenly show poor emissions performance.
`Figure 7 illustrates the response of the emissions rate (g/mi, not total emissions) to varying test
`lengths for an HEV and a conventional vehicle (HEV engine-on operation is indicated by gray sections
`on the graph). For a conventional vehicle, the initial start-up emissions are mitigated by a larger and
`larger proportion of cleaner emissions from hot engine operation; the g/mi rate over the whole cycle
`tapers off, which is a changing, but more or less predictable, continuous trend. In contrast, the emissions
`response from one possible ZEV-Capable HEV with engine on/off operation starts with zero emissions
`until a distance where the engine starts; then the line tapers off until engine shut-down, where it tapers
`off at a steeper rate. The next step increase is the second engine start-up.
`
`10
`
`BMW1030
`Page 11 of 15
`
`
`
`\ Conventional
`
`Test Length
`
`r
`
`Figure 7: Response in Emission Rates to Varying Test Lengths
`
`~
`
`Selecting one point on the HEV trend line does not necessarily represent a very good aggregate
`measure of the emissions over a wide spectrum of trip lengths; consequently, the results of the test can be
`very misleading. For instance, the emissions rates of an HEV during one particular trip may be
`especially high, but the following trip may actually operate with zero emissions (while the conventional
`vehicle will always produce emissions during every trip).
`
`PROPOSED HEV TEST C