# Steady State Speeds Load Determinations Using Electric Vehicle Power or Dynamometer Measurements on Conventional Vehicles.

INTRODUCTION

On a test track, a vehicle's propulsion electrical power flow can be used to quantify a vehicle's total parasitic road drags (both a vehicle's mechanical drags and ambient wind drags combined or total road drags) by measuring an HEV's motor electrical power flows during steady state (SS) speeds. On a chassis dynamometer the vehicle mechanical drags of the vehicle itself can be measured by using the dynamometer to motor the vehicle at steady state speeds and then recording the force required at each speed. The same may be accomplished using the vehicle propulsion motor power flows.

In the vehicle emissions field, coasting vehicles on a track and recording its velocity change with time has historically been used to determine vehicle total road drags or Track Target loads. The vehicle deceleration rates and inertial mass are used to calculate the forces (ambient and mechanical) that slow the vehicle down on the road. Society of Automotive Engineers (SAE) Standard J2263 Road Load Measurement Using Onboard Anemometry and Coastdown Techniques is the most commonly accepted standard that delineates this process. The J2263 procedure requires that coastdowns be performed on a track using runs in each direction, that are paired to average out differences which can be associated with the opposing directions of each run pair. Forces are then calculated and graphed verses vehicle speed. A quadratic equation

(F=A+B*V+C*[V.sup.2]) (1)

of force (F) as a function of velocity (V) is then calculated to fit the points. The major variability contributors to the force calculations are vehicle mechanical drive systems and ambient influences, such that there is no one singular, quadratic equation that can repeatedly describe the forces that act to slow down vehicles under any road conditions. A vehicle will have instead families of drag curves acting on it, depending on conditions. Because of these test measurement variations, SAE J2263 requires five paired runs to be used to calculate an average quadratic total road drag equation for a vehicle. This average or mean of all these track coastdowns is used as a singular estimate of total road drag forces needed to be emulated when testing a specific vehicle on a chassis dynamometer.

STUDY DESIGNS

The vehicle total road drag RLD, on a test track, requires the vehicle to obtain a top speed and then to have the vehicle transmission put in neutral. The vehicle is then allowed to coastdown to a low enough speed to cover the operating range of the vehicle in a dynamometer test. Conventional vehicles normally have transmissions with mechanical neutral modes. When a conventional vehicle transmission is placed in neutral the vehicle wheels are mechanically separated from the engine and portions of its drive train components.

The vehicle Track RLD produces a total Track Target quadratic equation that models the sum of windage, vehicle drive train mechanical and tire drags associated with the vehicle on the road. On a dynamometer RLD, the vehicle is allowed to coastdown repeatedly.

The dynamometer iterates through multiple coastdowns, changing the drags it asserts on each coastdown, comparing the times achieved to those calculated from the vehicle Track Target RLD equation for the vehicle. When the vehicle repeats times that match the Track Target RLD equation calculated times, within acceptance limits, the dynamometer set load associated with those replicate runs, expressed as a quadratic equation, is used as dynamometer set coefficients for the vehicle test cell testing.

The vehicle's mechanical drags are assumed to be the same on the dynamometer as when on the road. The sum of dynamometer load and vehicle's own mechanical drag values theoretically emulate the total road drags that the vehicle experiences on the road. When coasting a vehicle down on a dynamometer both the vehicle mechanical drive train drags and the dynamometer set drags act in conjunction to slow the vehicle down. Matching the speed interval times from the Track RLD is considered proof that the dynamometer has correctly simulated the vehicle's track total road drags. The dynamometer can then subtract the equation for the loads that it set, during the coastdown, from the overall Track Target equation to determine what the vehicle's mechanical drag loads are.

HEVs and Battery Electric Vehicles (BEVs) may not have power trains and/or transmissions that allow the vehicles to have true neutral modes. This can be for a number of reasons; multiple propulsion engines (internal combustion, electric and hydraulic motors), use of planetary gearing systems, single direct drive or multiple direct drive motors to name a few.

Manufacturers have used different methods to perform track RLDs on HEV and BEV vehicles. Vehicles have been modified to install a mechanical clutch just for the purpose of track RLDs and laboratory testing. Vehicles have also had their software modified to simulate types of "virtual" neutrals. Either way, RLDs performed by modifying the vehicle allows for the possibility that the RLD forces, after modifications, may not be the same as the forces exerted on the production model of the vehicle on the road.

The technology in HEV and BEV vehicles now presents the possibility to quantify vehicle mechanical drags and to do it with unmodified test vehicles under actual road conditions. The power used by an electrical motor is directly proportional to the load exerted on the drive motor or motors. The precision needed in the power control design of current HEV and BEV vehicles may allow us to measure motor amperage, voltage and subsequently the power that a vehicle uses to overcome the total road drags (i.e. ambient windage and vehicle mechanical drags) extremely accurately. This measured power should be able to be used with precisely measured vehicle speed to quantify the force exerted on a vehicle at any specific speed.

Using a vehicle held at steady state speeds to evaluate vehicle mechanical drags associated with vehicle drive trains is not a new practice. Automotive manufacturers, dynamometer manufacturers and Environmental Protection Agency (EPA) have used SS speeds to quantify a vehicle's total road drags for some time. On a test track, vehicles can be instrumented with either drive shaft torque meters or torque wheels, in order to record vehicle parasitic loads, as in separate studies performed by both Argonne Labs' Mike Duoba and EPA heavy duty truck regulators. Recorded torque values have then been analyzed verses specific steady state speeds to model the total road drags that a vehicle works against on the road, with a quadratic equation.

The various drag equations can be used in more than one way. As stated above, RLD coastdowns can be used on a chassis dynamometer, in a trial and error sequence of iterations to determine the dynamometer load curve necessary to match track time data. Knowing the dynamometer set coefficients, which result in acceptable CD time intervals, and subtracting them from the Track Target coefficients subsequently produces a vehicle's mechanical drags. By reversing this calculation, a chassis dynamometer can motor the vehicle at set speeds while the dynamometer measured forces are recorded. This data is then analyzed to determine the specific mechanical drag coefficients associated the vehicle. Once the vehicle's mechanical drags are determined they can be subtracted from a vehicle's Track Target RLD quadratic equation to calculate the necessary dynamometer set load needed to simulate the Track Target RLD quadratic equation.

Actual Programs Performed
```Canadian Vehicle Steady State Speed verses J2264 Coastdown Vehicle Loss
Measurement

Manufacturer   Vehicle Descriptions
Chrysler Compass 3500 Pound Mass
Chrysler       (lbm) Equivalent Test Weight (ETW)
Spark Ignition (SI)
Chrysler       Patriot 3625 lbm ETW, SI
Chrysler RAM 2500 Compression
Chrysler       Ignition (CI) Diesel 8000 lbm ETW
Ford           Escape Hybrid #1 4250 lbm ETW
Ford           Escape Hybrid #2 5000 lbm ETW
Honda          CR-V 4000 lbm ETW SI
Toyota         RAV4 4000 lbm ETW SI
Toyota         Highlander Hybrid 5000 lbm ETW
Toyota         Highlander 4750 lbm ETW SI
```

The EPA National Vehicle Fuels and Emissions Laboratory (NVFEL) has investigated using SS speed to measure vehicle mechanical drags (Vehicle Losses) compared to traditional J2264 "Chassis Dynamometer Simulation of Road Load Using Coastdown Techniques" coastdown RLD calculated vehicle mechanical drags since 2004. In 2009 NVFEL agreed to test nine four-wheel drive (4WD) vehicles for Environment Canada on a 4WD dynamometer. The Canadian government emissions testing program purchases vehicles for that use. They are mainstream vehicles that consumers purchase. No special care was taken with the vehicles used in this testing. They were shipped to the NVFEL and each tested multiple times over a couple of months' period of time. NVFEL took this opportunity to perform vehicle mechanical drag comparisons, between the two vehicle mechanical drag evaluation methods (RLD and SS), on a fair cross section of 4WD vehicles. The group included those listed below.

After all emission tests, J2264 RLDs and SS speed Vehicle Loss tests were performed and the data collected.

HEV Electric Power Propulsion verses J2264 Coastdown Vehicle Loss Measurement

NVFEL then investigated using vehicle electric propulsion power at SS speeds to measure vehicle mechanical drags (Vehicle Losses) compared to both traditional J2263 "Road Load Measurement Using Onboard Anemometry and Coastdown Techniques" and J2264 "Chassis Dynamometer Simulation of Road Load Using Coastdown Techniques" coastdown RLD calculated vehicle mechanical drags.

A GM Volt Hybrid Electric Vehicle (HEV) 3750 lbm ETW was used for all the track and dynamometer steady state (SS) speed electric propulsion power measurements.

Track Data Generation

The primary Volt electrical measurements, during the track runs of the SS Method, employed a Hioki Model 3390 Power Meter using their 9279 induction clamp. The Hioki Power Meter satisfies SAE J1711 "Recommended Practice for Measuring the Exhaust Emissions and Fuel Economy of Hybrid-Electric Vehicles, Including Plug-in Hybrid Vehicles". The meter calibration was up to date and it was an instrument used for Compliance testing at the EPA National Vehicle Fuels and Vehicle Emissions Laboratory (NVFEL).

The current induction clamp was installed on the main battery feed to the motor inverter. Voltage was acquired from a voltage tap installed on the high-voltage side of the DC-DC converter. Vehicle Rechargeable Energy Storage System (RESS) or propulsion battery voltage was measured using DC voltage leads tapped off the vehicle battery cables. Vehicle electrical power usage was measured with the Hioki 3390 as well as calculated from its amperage and voltage measurements. Speeds were recorded using instrumentation called VBOX by RACELOGIC, a high accuracy GPS data logging system for vehicle on road testing. Signals from both of these instruments were integrated to evaluate power measurements with speeds.

It would be simpler and more economical to perform the SS method on any HEV or BEV vehicle if the vehicle's CAN Bus measured values could be used accurately enough for the measurements, instead of using external instrumentation. To determine the viability of using CAN Bus data for the SS method, the Volt's CAN Bus data were recorded independently of the Hioki meter during all the SS runs. The CAN Bus data was then aligned with the Hioki/VBOX data, using the vehicle speed traces recorded in each data set to match them up.

The electrical and speed values, both track and dynamometer, used in this study were selected from 30 to 60 second SS interval averages, during the most stable speed sub-sections. Because of this interval averaging, of the electrical and speed measurements, it was assumed that the VBOX and Hioki Power Meter different acquisition rates data would not introduce analysis errors due to any slight misalignment of the time stamped data streams.

Vehicle driven steady state speed values throughout the study were fairly impossible to reproduce exactly from each run to run performed. To facilitate more exact speed, force and electrical power load comparisons; all data were idealized from raw data into quadratic equations which were then used to calculate specific values using the exact same speeds for all the coastdown and steady state speed, power and load force evaluations.

The vehicle SS speed electrical values were generated on laps of an oval track at Chrysler Chelsea Proving Grounds. All SS speed data were obtained on the track straightaways and SS speeds intervals were matched in opposite direction pairs to average out ambient influences. The Track paired runs in each direction were performed five times and averaged to parallel the SAE J2263 procedure.

The J2263 values, in both directions, were generated with CDs on a straight away track at Chelsea constructed for performing the J2263 RLD procedure.

Once the track total road drag values for the GM Volt were determined with both the traditional coastdowns and the SS method, the vehicle was brought to EPA's NVFEL for the test cell portion of the study.

Dynamometer Data Generation

In this GM Volt testing, it was necessary to determine what vehicle mechanical drag values were included in the track determined J2263 RLD quadratic targets. The GM Volt performed RLDs using SAE J2264 "Chassis Dynamometer Simulation of Road Load Using Coastdown Techniques" procedure. Four separate consecutive runs of the SAE J2264 procedure RLDs were performed on the GM Volt, three of which were used to evaluate its vehicle mechanical drag variability on the dynamometer.

In order to conserve the charge in the vehicle battery, first the dynamometer motored the GM Volt to warm the vehicle tires/drive train components, then the first RLD coastdown was used as an additional vehicle warm up. The calculated vehicle mechanical drag values of the last three SAE J2264 procedure runs were averaged and the averaged vehicle mechanical drags were compared against vehicle mechanical drag values from each individual run's calculated vehicle mechanical drags.

Using the J2264 procedure, RLDs were performed, in both the Canadian and GM Volt portions. The vehicle mechanical drags were quantified using MAHA 48" 4 Wheel Drive Dynamometer "Vehicle Loss Test" mode software. The dynamometer motors the vehicle through SS speed steps. Each SS speed step was held for a set length of time, in this case 30-60 seconds. This is analogous to the SAE J2452 "Stepwise Coastdown Methodology for Measuring Tire Rolling Resistance", averaging tire drags at steady state speeds, from both accelerating and decelerating approaches, minimizing effects that may be attributable to either accelerating or decelerating to each speed state. We chose to analyze from 15 mph to 75 mph in 10 mph steps up and then back down again to be analogous to the SAE J2452 procedure. The Vehicle Loss software then records the vehicle mechanical drags that it had to work against at each steady state speed programmed and generates a quadratic curve model of the vehicle's mechanical drag losses. The steady state speed determined vehicle mechanical drag curve is analogous to the J2264 procedure vehicle mechanical drag curve calculated by subtracting the dynamometer set coefficients from the vehicle Track Target values. Reversing the calculation in the J2264 procedure, the vehicle's mechanical parasitic loss curve can be subtracted from the test vehicle's Track RLD in order to determine what dynamometer set load coefficients will need to be to simulate the vehicle's Track RLD results.

The dynamometer's Vehicle Loss function was employed to automatically generate the SS speed steps and calculate vehicle mechanical drag quadratics. But, in the GM Volt study, we chose to work with raw data whenever possible, including dynamometer measurement of the Volt mechanical drags. This enabled using exactly the same data acquisition/analysis on all data, Track or dynamometer, with Hioki Power Meter data logs at a 1 hertz rate. To simplify combining the data from the multiple sources (Hioki, GM Volt CAN Bus and dynamometer logs), the CAN Bus data (speed, amperage, voltage and power) and dynamometer data (speed and load at surface of the roller) were all averaged to 1 hertz values. Again, the analyses always used averaged values over the longest most stable steady state speeds periods that the files contained. After averaging the individual data point files (both 10 and 100 Hertz) to be 1 hertz values, we were unable to discern any impacts on the results, either in magnitude or response rates, when compared with the original file data. Converting and analyzing all the data, from multiple sources, using the same data frequency, allowed us to more easily match up vehicle speeds contained in the various data files, for the analyses. Combining all the data files also allowed us to use the same curve fitting software (Excel) for all amperage, voltage, power, speed and pound force (lbf) load analyses. It was hoped that this would minimize any variability that may have been associated with using different curve fitting methods.

The next task was to try and incorporate the Test Track measured vehicle propulsion power during steady state speed measurements. These power measurement values were used as targets to determine what dynamometer set loads resulted in matching the real world loads experienced by the Volt on the test track. The dynamometer's Automatic Torque Regulation (ATR) constant load at surface of the roll setting function performed admirably for this portion of the study. The vehicle was set up on the dynamometer and held at constant speeds, by the driver, while the dynamometer load was increased until the vehicle power, measured by the Hioki, matched those that were recorded on the track for the same vehicle speed on Track.

CAN Bus data was obtained on the Track with an in house developed CAN Bus reader. CAN Bus data generated on the dynamometer testing used a DashDAQ-XL as a reader. The graphs of the Hioki and CAN Bus electrical data, on the Track runs, reflected each other with very linear correlation as can be seen in the bottom graph of vehicle CAN bus values verses Hioki 3390 measurements in Figure 1. Equivalent value spreads were also exhibited between the idealized SS Hioki and CAN Bus horsepower (HP) curves (5 runs each, see Figure 10).

During the dynamometer testing a DashDAQ-XL CAN Bus reader was used. This CAN Bus instrument provided more specific information than the one used on the track tests. The vehicle CAN Bus test data generated during the dynamometer runs provided propulsion powers of both Motor A and Motor B, in addition to amperage and voltage measurements of each, for total vehicle electrical power calculations. The vehicle coastdown data reviewed in this study consisted of speed verses time values from both Track coastdowns and dynamometer coastdowns.

The dynamometer ATR mode was used to exert constant loads on the vehicle. This allowed us to increase the load (load at surface of the roll) to the vehicle tires while holding the vehicle speed constant. When the readings on the Hioki 3390 Power Meter appeared to match the averaged values taken on the track at equivalent SS speeds the dynamometer measure forces were recorded.

DATA REVIEW AND ANALYSIS METHODS

During the oval track SS speed runs, the speed and power sampling periods, were kept to time interval windows of at least 30 second. Using DataDesk[R] allowed selecting contiguous sets of data with the most stable constant speeds in calculating average values of the variables monitored, an example of which is displayed below. (Figures 1 & 2)

The averaged values (both Hioki and CAN Bus data) run at SS speeds on an oval track were then graphed and ideal quadratic equations were generated for each of all the runs (Figure 3 & 4).

The results of the coastdown RLDs performed at Chelsea Proving grounds were reviewed and analyzed with the same methods to try and keep the results (Figure 5) as comparable as possible.

The next task was to try and replicate the vehicle total road drag forces, that the GM Volt exhibited at Chelsea Proving Grounds, on a dynamometer. Three different methods were used to determine dynamometer sets that could be used to match the test track CDs.

In the first method, three sets of the SAE J2264 "Chassis Dynamometer Simulation of Road Load Using Coastdown Techniques" procedure were used. Three sets were performed in order to examine what type of vehicle RLD variability can be expected when using the most commonly accepted traditional method of dynamometer Road Load Derivations. The dynamometer iterates through coastdowns changing the dynamometer set loads until the speed interval times predicted by the Track Target equation

([A.sub.track] + [B.sub.track]*V + [C.sub.track]* [V.sup.2]) (2)

are matched to times equal to an accepted tolerance (+/- 2.2 lbf in this study). Once a dynamometer setting

([A.sub.set] + [B.sub.set]*V + [C.sub.set]*[V.sup.2]) (3)

resulted in satisfying the tolerance, on three consecutive CDs, the run was complete. Subtracting a final dynamometer set equation from the Track Target equation results in a vehicle mechanical drag estimate.

([A.sub.veh] + [B.sub.veh]*V + [C.sub.veh]*[V.sup.2]) (4)

In order to get as robust an estimate for vehicle mechanical drags on the dynamometer for this study as possible, an averaged quadratic equation result from the three complete J2264 runs, was used as the overall J2264 dynamometer estimated vehicle mechanical drag (Equation 4)

As a second method, the vehicle was motored by the dynamometer through a series of SS speed steps. A vehicle mechanical loss quadratic equation (Equation 4) is generated from dynamometer motored SS speeds and the measured loads at the surface of the roller.

During these SS speeds. The vehicle mechanical loss quadratic equation, fit to these load values, can then be subtracted from the vehicle's Track Target CD determined overall total road drag (Equation 2) (vehicle mechanical plus windage drags). The difference between these two equations is then what the dynamometer set coefficients need to be (Equation 3) in order to simulate the total vehicle road load in the real world.

The EPA has investigated the equivalence between the SAE J2264 procedure vehicle mechanical drags, determined with CDs, and dynamometer SS speed measured vehicle mechanical drags since 2004. The two techniques have yielded averaged results within +/- 2% of each other and generally within 2 pound of force of each other. The same appeared true in this study (Figures 6 & 11).

For a third method, dynamometer load set coefficients were derived by holding the vehicle at SS speeds and loading the dynamometer until the Hioki meter readings matched ones which had been quantified during test track SS speeds. These loads were used to generate quadratic dynamometer load set coefficients (Equation 3). As in the first method, when an average dynamometer load set equation is subtracted from the vehicle Track Target quadratic equation the result is an estimate of a vehicle's mechanical drag quadratic equation (Equation 4).

Actual Observations

Of the nine 4WD vehicles tested, seven vehicles had the J2264 RLDs calculated vehicle mechanical drags, below 65 MPH, agree within +/- 5 lbf to SS speed measured ones. Five of those were within +/- 2 lbf from 10 to 70 MPH. NVFEL has also performed comparisons between these two methods, in testing their five 4WD dynamometers for acceptance, and the results were always within +/- 2 lbf. The results of all this testing strongly supports an equivalence between the steady state and traditional coastdown methods.

Two Ford Hybrids exhibited greater than +/- 2 lbf difference between traditional coastdown RLD and SS procedures (Figure 6).

One vehicle at speeds above 25 MPH and the other above 55 MPH. Graphs of the vehicles' deltas, between SAE J2264 and SS speed procedures, were roughly parallel. Both Fords with hybrid electric drives and trans axles exhibited the differences with the other vehicles. This may be worth further investigation and study.

Again no special care was taken on any of these vehicles, to initially ensure that there weren't brake or other anomalous mechanical drag forces present. Nor was time spent to investigate individual vehicles afterwards, so possible causes are open to conjecture. Testing comparisons on vehicles of this type may be helpful in deciding if CDs are the best methods to quantify vehicles drags on all vehicles.

Two of the nine vehicles exhibited excessive differences between the two methods (Figure 7), a Toyota SI and a Chrysler RAM 2500 CI (diesel) vehicle. The two outliers of the nine Canadian test vehicle results follow:

Both of these vehicles exhibited SS curves that were anomalous from the other vehicles. The curves for these vehicles show repeatable mechanical drags (as exemplified by the dyno set force curve graphs) during SS speed runs, which are not apparent in the traditional CDs. (Figures 8 & 9)

The fact that the Chrysler RAM 2500 diesel exhibited extreme variation on its CDs may indicate something correctable in the vehicle's mechanical drags (brake or other incorrectly dragging mechanical parts). This, again, may be a positive aspect of the SS speed method, to exhibit vehicle mechanical drag anomalies that are repeatable and may be correctable, yet may not be seen in traditional CDs. The 10 MPH load peaks, correctable or not, must influence both the J2263 and J2264 RLD curves.

Hybrid Electric Vehicle Track/Dynamometer Testing

The equation idealized data from Track Hioki, CAN Bus and RLD coastdowns are displayed on the same graph below (Figure 10).

In Figure 10, the idealized coastdown vehicle total road drag HP curves, from the SAE J2263 traditional coastdown RLDs, appear to lay right on top of the vehicle total drag curves calculated from Volt CAN Bus data, exhibiting the same overall data values and spreads between the two. The Hioki measured vehicle total road drag curves show data spreads very similar to the coastdown RLD and CAN Bus curves but exhibit 10% - 15% higher biases to vehicle total road drags on J2263 and CAN bus curves. A couple of possible explanations of differences between the J2263 Track CDs and SS speed Hioki meter, in this study, may be attributable to motor inefficiencies and/or vehicle accessory loads for which the Hioki data (total battery amperage exchanges) were not corrected.

The GM Volt's mechanical drag quadratic equations calculated from runs with dynamometer set coefficients from all three techniques (J2264, Track Hioki SS speed power and dynamometer driven/measured J2452 SS speed technique) are displayed (Figure 11) below. All three methods appear to be able to exhibit results within the +/- 2 lbf of each other that we have seen between traditional J2264 and dynamometer motored steady state speed vehicle mechanical drag determinations.

On quadratic curves fit to data for each of the J2264 runs on the GM Volt, run to run differences were less than 0.3 lbf from the quadratic curve of the average values of the three runs. On quadratic curves for all the steady state runs on the GM Volt, run to run differences were also less than 0.3 lbf from the curve of averages of the three runs data.

The CAN Bus total electrical power measurements tracked very well with the Hioki 3390 Power Meter values taken during vehicle driven dynamometer SS speeds. When idealized values from three runs CAN Bus and Hioki values were compared, 20 values of 21 from 10 mph to 70 mph were within 1 kilowatt (kW) (Figure 12).

Finally, GM Volt vehicle CDs were then performed on the dynamometer, using dynamometer coefficients derived using Hioki Meter results from the test track and dynamometer SS speed runs. The results, in the three CDs using the Power meter derived dynamometer set coefficients, were within +/- 3 lbf of the Track Target curve, which was not corrected for wind, for all cases and the curves are displayed in Figure 13.

CONCLUSIONS

Testing on 4WD vehicles supports that evaluating vehicle loads using steady state vehicle speeds can give results acceptable to SAE J2263 and J2264 CD procedures.

For simplicity, data corrections for ambient conditions were not employed in this study. Using all the data corrections recommended in SAE J2263, to both the SS speed data and matching J2263 Track data, would have to improve agreement between the two methods.

Dynamometer data generated on J2264 data were not any more or less variable than the SS speed data.

Steady State speed quantification of vehicle mechanical drags may reveal anomalous vehicle mechanical drags which may not be apparent in traditional CD methods.

Improved results may be possible on vehicles given professional mechanical care to ensure that they are in manufacturer recommended working condition before any load evaluation testing begins. Drags resulting from improper maintenance or mechanical operation should be eliminated.

Track data generated using J2263 Track CDs and SS speed results from the Volt CAN bus data exhibited very good correlation.

SS speed on electric vehicles could be used to verify RLD results on vehicles modified because they do not have true neutral gears.

Hioki Power Meter measurements appear capable of reproducing Track Target loads on a dynamometer correctly.

The differences between the J2263 Track CDs and SS speed Hioki meter, in this study, may be attributable to motor inefficiencies and vehicle accessory loads.

It appears that GM Volt CAN Bus values can be accurate enough to quantify vehicle track loads at SS speeds. If initially verified against traceable independent standards, CAN Bus readings should be able to be used for SS vehicle load evaluations.

The SS speed method could be used for quantifying HEV propulsion efficiencies.

RECOMMENDATIONS

Additional testing in this area is warranted. Monitoring electrical power usage directly from propulsion motor cables or CAN Bus readers (especially individual motors separately) show very promising results for Vehicle Emissions testing RLD and Research & Development.

Adjusting RESS battery amperage numbers, for known motor efficiencies, could improve the steady state external instrumentation power measurement agreement with J2263 CD Track Target quadratic equation coefficients.

If it is possible, monitoring any electrical energy going to accessories would allow for correction of a gross amperage out measurement as used in this study.

Vehicles used in studies of this type should be examined and maintained by qualified mechanics to ensure that they are operating mechanically correctly.

The three vehicles exhibiting anomalous results should have been investigated and established to be in proper working condition by trained mechanics.

REFERENCES

[1.] SAE International Surface Vehicle Recommended Practice, "Road Load Measurement Using Onboard Anemometry and Coastdown Techniques," SAE Standard J2263, Rev. Dec. 2008.

[2.] SAE International Surface Vehicle Recommended Practice, "Chassis Dynamometer Simulation of Road Load Using Coastdown Techniques," SAE Standard J2264, Iss. Apr. 1995.

[3.] SAE International Surface Vehicle Recommended Practice, "Stepwise Coastdown Methodology for Measuring Tire Rolling Resistance," SAE Standard J2452, Iss. June 1999.

[4.] MAHA Emissions Chassis Dynamometer AIP-ECDM 48M/4x4 SN 421

CONTACT INFORMATION

Carl Paulina

United States Environmental Protection Agency

Office of Transportation and Air Quality

National Vehicle and Fuel Emission Laboratory

paulina.carl@epa.gov

ACKNOWLEDGEMENTS

The authors thank the Chrysler Proving Grounds personnel and Jim Rollins in particular for use of their facilities and assistance. NVFEL personnel. Manish Patel and Scott Wilson performed all the testing on the Canadian 4WD vehicles. EPA Compliance Division engineer Jim Snyder for the use of the RACELOGIC VBOX instrumentation and advice. A special thanks to Ford Motor engineer and chairman of the SAE Light Duty Vehicle Performance and Economy Measure Committee Jeff Glodich, one of the originators and collaborators on this vehicle load Steady State speed method concept.

DEFINITIONS/ABBREVIATIONS

4WD - four wheel drive

ATR - Automatic Torque Regulation

BEV - battery electric vehicle

CD - vehicle coastdown

EPA - Environmental Protection Agency

ETW - equivalent test weight

F - force

HEV - hybrid electric vehicle

HP - Horsepower

IW - inertia weight

lbf - pound force

lbm - pounds mass

MPH - miles per hour

NVFEL - National Vehicle & Fuels Emissions Laboratory

RESS - Rechargeable Energy Storage System

SI - spark ignition

V - velocity

Carl Paulina, Dan McBryde, and Mike Matthews

US EPA - NVFEL

doi:10.4271/2017-01-0990