See coast down testing documentation.

October 14, 2011
SmartTruck Coastdown Testing Program
UT-1b Configuration
1 The SmartTruck Heavy Vehicle Coastdown Test Protocol
The SmartTruck coastdown protocol uses a combination of high speed test runs with
coastdown from 70 mph to 55 mph and low speed test runs with coastdowns from 40
mph to 15 mph to obtain the required high speed drag data and the value Crr o with
which to correct the total drag. Figure 1 below shows that the accelerate-coastdown
distance for the high speed runs is just over 5,000 feet for a combination weight of
36,500 lbs.
Figure 1
There are many facilities available with this length and adequate turn around tracks.
SmartTruck has tested at Michelin’s Proving Grounds Track 9 (available for rent to the
public) and an inactive runway at the Donald Center in Greenville to perform these
tests. This allows local, cost effective testing to be done on many configurations. Figure
2, Figure 3 and Figure 4 below show actual raw data from the SmartTruck data system
for a single configuration run.
Figure 2 - Low Speed Runs
2
Figure 3 – High Speed Runs
Figure 4 - High Speed Runs
The blue lines are the vehicle speed measured with both driveshaft and front wheel
sensors. The red lines are truck airspeed data from a calibrated pitot static system on
board the tractor.
While the airspeed system is not strictly needed for good Cd
measurement as long as the winds are low and consistent, it is needed to measure the
time variant Cd during any given run. SmartTruck uses the time variant Cd to get
average Cd, and to see if our aerodynamic modifications reduce or increase the
frequency or magnitude of Cd variations. We also use the airspeed system data to
disqualify a run with excessive gusting or yaw within in a run. We measure the yaw
angle with our data system directly but again this is not strictly necessary for good
average Cd data if a good weather station is used as is required by both protocols.
Airspeed data contains a significant high frequency content that is related to cab
3
vibration not gusting. This must be removed from the data to obtain good time variant
Cd information. Figure 5 below shows the raw signal (blue) and the filtered signal (red)
that is ultimately used in the calculations.
Figure 5
Figure 6 shows results of regression analysis of the low speed runs used to obtain Crr0
for removal of the rolling resistance and friction from the total retarding force to get the
aerodynamic drag force. For this case, the baseline truck, the average F total at zero
speed is -173.57 lbs giving CRR0 = 173.57/36,541 = .00475
4
Figure 6 - Low Speed Run Results
5
Figure 7 shows typical Cd vs. time data on one of our Track 9 test runs.
Figure 7
The blue line is the time accurate Cd, the red line is the average Cd. The unsteadiness
in the time accurate data is due to variations in Cd due to the unsteady nature of the
wake flows on the rear of the trailer, the gap, and the underside. This is a real effect
and not and artifice of the testing technique.
The average Cd’s for all runs are averaged for a final Cd value for the configuration.
Average Cd’s are also checked for too great a run to run variance in which case that run
is eliminated and repeated.
SmartTruck has tested over 200 configurations using this protocol. We test our baseline
configuration at every test and several times during a test day and consistently get
accurate and repeatable results both within a test day and between tests going back
over two years.
6
The following is a detailed test report for coastdown testing of our UT1b variant using
our protocol for your review.
2 Approach
This testing program was done in accordance with proven coastdown testing
techniques. To further facilitate proper implementation of the protocols, a consistent
Kentucky 53 foot dry van trailer and Navistar 2009 model year ProStar Tractors were
used. This combination remained consistent throughout testing.
The test truck was equipped with state of the art data acquisition systems. These
systems have 60 available channels to monitor and record a wide variety of vehicle
systems and effects, including, but not limited to:

True air speed via pitot tube and calibrated static reference pressure

Both driver and passenger wheel speeds

Drive shaft rpm

Engine rpm

Yaw angle/wind direction

Steering input

Vehicle lap times
Weather was monitored by a Davis Vantage Vue weather station, located halfway down
the track, to provide data as close to what the truck was exposed to as possible.
3 Test Procedure
Prior to each testing segment the truck operated on the track for a one hour warm-up.
Following the warm-up phase, two separate coastdown runs would be completed. The
first run was comprised of one low speed lap followed by two high speed laps. The
second run consisted of two high laps.
After each run a pit stop was preformed, where engineers would:

Download both Stack data system data and Cummins engine data.

Review of coastdown data to ensure integrity.
7

Tire pressures were checked to ensure they were not warming up nor cooling
down.

Tractor check list was performed to ensure it was still properly warmed up and in
proper working condition.

Weather station data was downloaded and check to ensure good weather
conditions.
4 Vehicle Preparation

All vehicle axles were aligned to manufacturer’s specifications. Tractor and trailer
axle bearing and brake adjustments were made at this time.

The tractor trailer gap was set in a commonly used long haul configuration.
Specifically, the King Pin location was set so that the back of the cab to the front of
the trailer gap was 50 inches.

The King Pin location of both trailers was set to the tenth (of a total of fourteen)
adjusting hole, forward of the rearward most adjustment point.

The rear trailer slider was set to 40 feet.
5 Weight
Fuel consumption for each vehicle was measured for each run completed.
Consumption, measured in pounds, was determined by reading the total fuel used from
the Cummins engine data and calculating the difference from the previous run. Weight
for each kit configuration was also accounted for. These weights are needed to ensure
SmartTruck knows the weight of the test vehicle prior to each coastdown run.
8
6 Vehicle and Equipment Specifications
Tractor
Trailer
Unit #
C13602
D14272
Make
Navistar
Kentucky
Model
Pro Star 6.4
N/A
V.I.N.
3HSCUAPR49N144465
1KKVA53279L228141
ISX 435 ST
N/A
31,020
N/A
Tires-Steer
Goodyear G395LH5
Fuel Saver 295/75R
22.5 (110psi)
N/A
Tires-Drive/Trailer
Goodyear G305LH5
Fuel Saver 295/75R
22.5 (110psi)
Bridgestone V-Steel
Rib Radial R195 11R
22.5 (105psi)
RTD 16910B-DM3
N/A
Engine
Start Odometer
Transmission
Table 1 - Tractor-Trailer Information
7 Description of Test Facility
Testing was conducted in Laurens, South Carolina on the Michelin Tires Laurens
Proving Grounds (LPG). LPG is a state of the art testing facility with a total of nine
unique tracks including: a main test track, road course, wet handling, gravel endurance,
off road inclines, heavy truck loop, noise, vehicle dynamics and drift/pull.
Figure 8- LPG Facility Map
9
SmartTruck currently takes advantage of LPG’s Track 9, Drift/Pull. This track is a 4,800
foot straightaway with turnaround loops on either end for a total length of 1.25 miles.
The track width is 40 feet in the turnarounds and 80 feet in the straightaway. The
surface of the track is asphalt with a surface texture (Macro/Micro) of smooth/rough.
Track 9 also has a near perfect flatness over the straightaway length with an
International Roughness Index (IRI) of 37.4 in/mile.
Figure 9 - Track 9, Drift/Pull
8 Test Configuration
Following the conclusion of all baseline testing and calculations, the test truck was
outfitted with SmartTruck’s UT-1b Trailer UnderTray system.
consists of:
1. Forward UnderTray.
2. Integrated Axle Sled.
3. Aerodynamic Rain Gutter.
Figure 10 - Trailer Equipped With UT-1b
10
This configuration
9 Rolling Resistance
Rolling resistance at zero speed was measured for each configuration from the low
speed runs and the actual RR curve was Crr = Crr
zero
+ (5*10-7)*V2. This was done for
each configuration.
10 Drag Calculation Equations
Daero = (Wc/g)*dVwheelspeed/dT – Crr*W (V in f/s)
Cd = Daero/Aref/.5*RHO*V2airspeed
Crr = Crr0 + (5*10-7)* V2wheelspeed (V in mph)
Where:
Wc is vehicle weight in lbs (which includes the inertial effects of the wheels)
W is vehicle weight in lbs
Aref – 97.2 Square feet
RHO is measured air density in slug –ft/sec2
Vairspeed is the measured airspeed in f/s
Vwheelspeed is the measured vehicle speed in mph
Crr is the coefficient of rolling resistance
Crr0 is the coefficient of rolling resistance at zero speed
11 Aerodynamic Drag Results
Baseline
UT-1
UT-1b
Crr
0.00475
0.00480
0.00474
Cd
0.58750
0.53458
0.53300
% Aero Improvement
-8.96%
9.28%
Table 2 - Drag Results
12 Weather Data
Baseline
Air Temperature (ºF)
Air Density (slug/
)
UT-1b
Min.
Max.
Avg.
Min.
Max.
Avg.
64.9
73.1
68.8
76.8
77.8
77.2
0.002241
0.002277
0.002260
0.002221
0.002225
0.002224
1 Min Average Wind Speed (mph)
1
7
3
3
7
5
High Wind Speed (mph)
2
12
5
4
10
7
Table 3 - Weather Data
11
Attachment 1
General Discussion of EPA’s Modified Protocol based on
SAE J1263 and J2263 Coastdown Protocols
1 Discussion of Coastdown Testing For Heavy Vehicles
EPA's Modified Protocol based on SAE J1321 and J2263 coastdown protocol has been
suggested for testing of Class eight trucks to qualify aerodynamic devices on the tractor
and the trailer.
Our experience has been, after testing more than 200 different
aerodynamic configurations, is that there are several issues with the suggested protocol
which make it virtually impossible to archive accurate results and very difficult and
expensive to perform the testing.
2 Issues in Heavy Truck Testing using EPA's Modified Protocol
based on SAE J1321 and J2263
2.1 Issue 1 – 70 mph to 17 mph Coastdown Interval
This coastdown interval is required for the data reduction technique spelled out in the
protocol to work accurately i.e. obtaining the zero velocity drag force for rolling
resistance correction.
SAE J1321 and J2263 protocols were developed for light
vehicles (basically automobiles and light trucks) that could accelerate to 70 mph and
then coastdown to less than 17 mph in a reasonable distance (about 6000 ft) due to
high drag to weight ratio typical of cars and light trucks. There are many facilities that
are available that are long enough for this test with cars and light trucks. However, a
Class 8 tractor-trailer combination, completely unloaded, weighs in the order of 36,000
pounds. It’s power to weight and drag to weight is a fraction of a car or light truck.
Consequently the total distance required to properly run EPA's Modified Protocol based
on SAE J1321 and J2263 is typically greater than 13,000 feet. See Figure 11.
12
Figure 11 - Calibrated Truck Model Results for Class 8 Accelerate and Coastdown Distance
Not many facilities offer this size track. SmartTruck has a Space Act agreement with
NASA to use their Space Shuttle runway (which is 18,000 feet in length) and we have
conducted multiple testing programs there using a coastdown of 70 mph to less than 15
mph. The Shuttle runway is active (Astronauts and their crews operate out of this
facility) and has heightened security so scheduling and operations are quite difficult.
Our experience is that this is a very expensive facility that few would take advantage of,
yet the existing protocols technically require this type of venue.
2.2 Issue 2 – Assumption That The Rolling Resistance and Friction Is
Constant i.e. Does Not Vary With Speed
Rolling resistance (and friction) is accounted for in EPA's Modified Protocol based on
SAE J1321 and J2263 by plotting the instantaneous total force calculated from the
measured dV/dT and vehicle weight versus velocity and then extrapolating it to zero
speed. Since the aerodynamic drag is zero at zero speed, the intersection represents
the rolling resistance and friction forces at zero speed. This force is then subtracted
from the total force to extract aerodynamic drag at the desired speed. Figure 12 below
is a typical curve of this sort from a SmartTruck test at the Kennedy Space Center.
13
Crro = -159/36050. = .0044
Figure 12
As can be seen the intercept with the y axis is at a retarding force of 159 pounds. This
divided by the weight gives a coefficient of rolling resistance (Crr) of 0.0044. This is
consistent with our experience with the tires used on our test trailer at zero speed.
However, if one uses data on Crr from the tire companies and literature one finds out
that Crr varies as the square of speed. Indeed our data for the tires we use and other
data on other test tires suggest that the coefficient of rolling resistance follows the
following formula:
Crr = Crro + (5x10-7)*V 2mph
When this formula is used for data reduction a much more accurate drag prediction is
calculated because the rolling resistance and friction drag are not constant and the
difference in rolling resistance at speed and the zero speed value gets added to the
“aerodynamic” drag value. Figure 13 below is again from our Kennedy testing and
shows the difference in the drag prediction when Crr is constant and when the formula
above is used.
14
Figure 13 – Kennedy Space Center Coastdown Results
The red line is the Cd predicted using the variable Crr, the blue is the Cd predicted
using the constant value of Crr=Crro.
The red line is nearly constant with speed and very closely agrees with the CFD
predicted value of Cd (0.59) as well as the Cd implied by our fuel mileage testing of this
configuration while the Cd predicted by EPA’s Modified Protocol is high (due to the
infusion of rolling resistance and friction drag in the aerodynamic drag levels) and
significantly variant with speed which is inconsistent with any other analysis of drag.
Errors in the relative drag levels using EPA’s Modified Protocol are of course smaller
then the absolute level error but still can be significant since the Crr error is constant.
As the aerodynamic drag is reduced the Crr error is a larger percent of the total
predicted drag level thus increasing the Cd level relative to a higher drag baseline.
Using a varying Crr is not prefect but errors in the Crr slope represent much smaller
differential errors than just assuming the slope is zero.
Again, light vehicles get away with this because of their higher aero drag to rolling
resistance ratio due to their lighter weight. In heavy vehicles the error is too great.
15
Attachment 2
Summary of Current SmartTruck UT System Configurations
Improvements at 65 mph
Smart Truck Product
Standard Trailer
SAE J1321
CD
Model
%MPG
Measured
0.5870
0.00%
Coast Down Results
CD
%MPG
Measured
Model
UT6(V2) w /Side Fairings
UT6 Base System (v2)
UT6 Base System (v1)
CFD Results (corrected)
CD Calculated
%MPG
& Corrected
Model
0.5875
0.00%
0.5870
0.00%
0.4980
10.20%
0.5115
8.40%
0.5215
7.40%
0.5269
6.58%
0.5280
6.50%
0.5280
6.50%
0.5260
6.80%
0.5330
6.10%
0.5275
6.30%
UT1 w/Side Fairings
0.5360
5.50%
0.5349
5.67%
0.5349
5.67%
UT1R (SIDES+ARG10B)
0.5430
4.80%
0.5450
4.56%
0.5450
4.56%
UT1 w/Integrated Sled and Side Fairings
UT1 w/Integrated Sled
16
November 30, 2011
SmartTruck Coastdown Testing Program
UT6v2 Configuration
1 Approach
This testing program was done in accordance with proven coastdown testing
techniques. To further facilitate proper implementation of the protocols, a consistent
Kentucky 53 foot dry van trailer and Navistar 2009 model year ProStar Tractors were
used. This combination remained consistent throughout testing.
The test truck was equipped with state of the art data acquisition systems. These
systems have 60 available channels to monitor and record a wide variety of vehicle
systems and effects, including, but not limited to:

True air speed via pitot tube and calibrated static reference pressure

Both driver and passenger wheel speeds

Drive shaft rpm

Engine rpm

Yaw angle/wind direction

Steering input

Vehicle lap times
Weather was monitored by a Davis Vantage Vue weather station, located halfway down
the track, to provide data as close to what the truck was exposed to as possible.
2 Test Procedure
Prior to each testing segment the truck operated on the track for a one hour warm-up.
Following the warm-up phase, two separate coastdown runs would be completed. The
first run was comprised of one low speed lap followed by two high speed laps. The
second run consisted of two high laps.
After each run a pit stop was preformed, where engineers would:

Download both Stack data system data and Cummins engine data.

Review of coastdown data to ensure integrity.

Tire pressures were checked to ensure they were not warming up nor cooling
down.

Tractor check list was performed to ensure it was still properly warmed up and in
proper working condition.

Weather station data was downloaded and check to ensure good weather
conditions.
3 Vehicle Preparation

All vehicle axles were aligned to manufacturer’s specifications. Tractor and trailer
axle bearing and brake adjustments were made at this time.

The tractor trailer gap was set in a commonly used long haul configuration.
Specifically, the King Pin location was set so that the back of the cab to the front of
the trailer gap was 50 inches.

The King Pin location of both trailers was set to the tenth (of a total of fourteen)
adjusting hole, forward of the rearward most adjustment point.

The rear trailer slider was set to 40 feet.
4 Weight
Fuel consumption for each vehicle was measured for each run completed.
Consumption, measured in pounds, was determined by reading the total fuel used from
the Cummins engine data and calculating the difference from the previous run. Weight
for each kit configuration was also accounted for. These weights are needed to ensure
SmartTruck knows the weight of the test vehicle prior to each coastdown run.
2
5 Vehicle and Equipment Specifications
Tractor
Trailer
Unit #
C13602
D14272
Make
Navistar
Kentucky
Model
Pro Star 6.4
N/A
V.I.N.
3HSCUAPR49N144465
1KKVA53279L228141
ISX 435 ST
N/A
29,560
N/A
Tires-Steer
Goodyear G395LH5
Fuel Saver 295/75R
22.5 (110psi)
N/A
Tires-Drive/Trailer
Goodyear G305LH5
Fuel Saver 295/75R
22.5 (110psi)
Bridgestone V-Steel
Rib Radial R195 11R
22.5 (105psi)
RTD 16910B-DM3
N/A
Engine
Start Odometer
Transmission
Table 1 - Tractor-Trailer Information
6 Description of Test Facility
Testing was conducted in Laurens, South Carolina on the Michelin Tires Laurens
Proving Grounds (LPG). LPG is a state of the art testing facility with a total of nine
unique tracks including: a main test track, road course, wet handling, gravel endurance,
off road inclines, heavy truck loop, noise, vehicle dynamics and drift/pull.
3
Figure 1- LPG Facility Map
SmartTruck currently takes advantage of LPG’s Track 9, Drift/Pull. This track is a 4,800
foot straightaway with turnaround loops on either end for a total length of 1.25 miles.
The track width is 40 feet in the turnarounds and 80 feet in the straightaway. The
surface of the track is asphalt with a surface texture (Macro/Micro) of smooth/rough.
Track 9 also has a near perfect flatness over the straightaway length with an
International Roughness Index (IRI) of 37.4 in/mile.
Figure 2 - Track 9, Drift/Pull
7 Test Configuration
Following the conclusion of all baseline testing and calculations, the test truck was
outfitted with SmartTruck’s UT6v2 Trailer UnderTray system.
consists of:
1. Forward UnderTray.
2. Integrated Axle Sled.
3. Rear Diffuser
4. Aerodynamic Rain Gutter.
4
This configuration
Figure 3 - Trailer Equipped With UT6v2
8 Rolling Resistance
Rolling resistance at zero speed was measured for each configuration from the low
speed runs and the actual RR curve was Crr = Crr
zero
+ (5*10-7)*V2. This was done for
each configuration.
9 Drag Calculation Equations
Daero = (Wc/g)*dVwheelspeed/dT – Crr*W (V in f/s)
Cd = Daero/Aref/.5*RHO*V2airspeed
Crr = Crr0 + (5*10-7)* V2wheelspeed (V in mph)
Where:
Wc is vehicle weight in lbs (which includes the inertial effects of the wheels)
W is vehicle weight in lbs
Aref – 97.2 Square feet
RHO is measured air density in slug –ft/sec2
Vairspeed is the measured airspeed in f/s
Vwheelspeed is the measured vehicle speed in mph
Crr is the coefficient of rolling resistance
5
Crr0 is the coefficient of rolling resistance at zero speed
10 Aerodynamic Drag Results
Baseline
UT6v1
UT6v2
Crr
.00470
.00470
.00470
Cd
.5875
.5280
.5215
% Aero Improvement
-10.1%
11.3%
Table 2 - Drag Results
11 Weather Data
Baseline
UT6v2
Min.
Max.
Avg.
Min.
Max.
Avg.
86.1
87.3
86.7
88.5
89.8
89.0
0.002186
0.002187
0.002187
0.002174
0.002181
0.002178
1 Min Average Wind Speed (mph)
1
3
2
2
7
4
High Wind Speed (mph)
2
5
4
3
11
8
Air Temperature (ºF)
Air Density (slug/
)
Table 3 - Weather Data
6
Attachment 1
CFD Summary - UT6v1 and UT6v2
CFD Comparison of UT6 Systems versus Baseline Trailer
BASELINE
Tractor
Trailer
0.28731
0.29920
UT6V1
as Tested
0.28757
0.24039
Total Tractor-Trailer
0.58650
0.00%
0.52796
9.98%
0.52169
11.05%
Front Wheels (and Fairing if applicable)
Rear Wheels (and Fairing if applicable)
Back
Top
Sides
Landing Gear
Suspension
Underside
ICC Bumper
Undertray System
Mud Flaps
0.03588
0.02199
0.15619
0.00961
0.02564
-0.00152
0.02199
0.00688
0.01658
0.00000
0.00596
0.03472
0.01925
0.10829
0.01743
0.02308
0.00133
0.01067
0.00636
0.00277
0.01649
0.00000
0.03477
0.02197
0.09977
0.01165
0.02558
-0.00149
0.01509
0.00770
0.00425
0.00831
0.00675
Total Trailer
0.29920
0.00000
0.24039
19.7%
0.23433
21.7%
% Improvement in drag
UT6V2
0.28737
0.23433
Trailer System Breakdown
% Improvement in drag
Table 4 - CFD Comparisons
7
0.18000
0.16000
0.14000
0.12000
0.10000
0.08000
0.06000
0.04000
0.02000
0.00000
-0.02000
Trailer System Drag Breakdown
BASELINE
UT6V1
UT6V2
Figure 4 - Drag Break Down
8
Attachment 2
Flow visualization - UT6v1 and UT6v2
Baseline
Figure 5- Flow Visualization from Baseline
9
UT6v1
Figure 6 - Flow Visualization from UT6v1
10
UT6v2
Figure 7 - Flow Visualization from UT6v2
11