Building a dynamometer test-stand

TU/e Masters Internship Report

Building a dynamometer test-stand

A.E. Dortland (534008)
22nd October 2007
Report no: DCT-2007-134

Supervisors

Prof. ir. A.A. Frank M. Sc. Ph.D.(UC Davis)

dr. P.A. Veenhuizen (TU/e)

Eindhoven University of Technology University of California, Davis

Department of Mechanical Engineering Department of Mechanical Engineering
Division Dynamical Systems Design Hybrid Electrical Vehicle Center
Master track Automotive Engineering Science
2
Contents

1 Summary 5

2 Introduction 7

3 ChallengeX 9
3.1 Well to wheel petrolium use . . . . . . . . . . . . . . . . . . . . . 9

4 Trinity 11
4.1 Driving modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
4.2 Electric motor and battery efficiency . . . . . . . . . . . . . . . . 12
4.3 CVT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.3.1 CVT chain . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.3.2 CVT hydraulics . . . . . . . . . . . . . . . . . . . . . . . . 15
4.3.3 CVT Control . . . . . . . . . . . . . . . . . . . . . . . . . 17

5 Prius engine 23
5.1 Why using ethanol . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5.2 Adjustments on the engine . . . . . . . . . . . . . . . . . . . . . 24

6 Dynamometer test stand 25

6.1 Brackets and shafts . . . . . . . . . . . . . . . . . . . . . . . . . . 25
6.1.1 Engine mounting brackets . . . . . . . . . . . . . . . . . . 27
6.1.2 Dynamometer mounting brackets . . . . . . . . . . . . . . 30
6.1.3 Dynamometer adaptor plate . . . . . . . . . . . . . . . . . 32
6.1.4 Flywheel adaptor shaft . . . . . . . . . . . . . . . . . . . . 33
6.2 Engine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

3
 6.2.1 Engine controls . . . . . . . . . . . . . . . . . . . . . . . . 35
6.2.2 Engine cooling . . . . . . . . . . . . . . . . . . . . . . . . 37
6.3 Dynamometer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
6.3.1 Hydraulic system of the dynamometer . . . . . . . . . . . 38
6.3.2 Controlling the dynamometer . . . . . . . . . . . . . . . . 40
6.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

7 Conclusion 41

A Symbols 43

B Derivation of the acceleration of the drive shaft 45

C Engine harness 47

D Technical drawings 49
D.1 Dynamometer adaptor plate part 1 . . . . . . . . . . . . . . . . . 49
D.2 Dynamometer adaptor plate part 2 . . . . . . . . . . . . . . . . . 51
D.3 Flywheel adaptor shaft . . . . . . . . . . . . . . . . . . . . . . . . 52
D.4 Encoder adaptor plate . . . . . . . . . . . . . . . . . . . . . . . . 53
D.5 Front engine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

4
Chapter 1

Summary

The mechanical engineering department of the University of California Davis

has a very progressive team working on plug-in hybrid vehicles. For several
years they have been participating in different competitions for hybrid vehi-
cles. The last three years they were re-engineering a Chevrolet Equinox for
the General Motors competition ’Challenge X’. Last year’s competition week
was a good week for UC Davis. The car showed a much better fuel economy
while improving the performance. A very important part of the car is a custom
built Continuous Variable Transmission. Professor Frank invented a new con-
trol strategy for this CVT. A simulink model shows the main idea behind this
control strategy. Another important part of the car is the engine. The team,
Team Fate, decided to use a Toyota Prius engine (Atkinson cycle) and to run
this engine on E85(ethanol). This engine helps to build a more sustainable car.
However the fuel economy is improved the engine was not tuned to burn E85.
The most important purpose of the powertrain and its control strategy is to
run the engine on the Ideal Operating Line. To achieve this some knowledge
about the engine which burns ethanol is necessary. A dynamometer stand for
the engine to obtain an engine map was built. Unfortunately this is not finished
yet but a lot of work is already been done.

5
6
Chapter 2

Introduction

For several years the mechanical engineering department of the University of

California Davis have been building hybrid electric vehicles. Mainly volunteers
are working in Team Fate. The team, lead by Professor Frank, builds cars with
much better fuel economy than conventional vehicles. Last year was the third
year Team Fate participated in a General Motors hybrid vehicle competition
called Challenge X. The last weeks before the assessment in Detroit a lot of
work was done on the car to get it operational. After the competition the
team started optimizing the car. This report first describes, in chapter three,
what this competition is about and reveals the own goals of the team. Chapter
four is about the car involved in this competition. The drivetrain topology will
extensively be explained. Different aspects of the transmission, a custom built
CVT and heart of the drivetrain, will be discussed. A control strategy for this
CVT, invented by Professor Frank, is implemented in Matlab. Some simulations
are done and results will be shown. Chapter five is about the used engine and
fuel: A prius engine which burns ethanol. The last chapter is about the actual
assignment of this internship: Building a test stand for the mentioned engine.
This test stand should provide data about the characteristics of the engine
burning ethanol. During the internship the car is extensively tested by GM in
Detroit. Some good results were achieved but the engine was not yet tuned
for ethanol so there is still room to increase fuel economy of the car. This test
stand is the first step to make this possible.

7
8
Chapter 3

ChallengeX

ChallengeX is the name of the competition where UC Davis is involved in

together with 16 other American universities. The challenge is to re-engineer
a General Motors Chevrolet Equinox to obtain energy sustainability, reduce
emissions and greenhouse gases while maintaining or exceeding the vehicle’s
utility and performance. The hybrid electric vehicle competition holds 4 years:
The first year the teams had to focus on modeling, simulation and testing of
the vehicle powertrain and vehicle subsystems selected by the university. UCD
decided to build a pre-transmission parallel plug-in hybrid. The judges were
positive about the work done by Team Fate, so GM donated a car to UCD to
apply their plan. Year 2 and 3 required the teams to integrate their powertrain
and subsystems in the donated GM Equinox. After each year the teams have
to come together to present their cars to a panel of judges. Several events
will test fuel economy, emissions, vehicle utility and performance. Year 2 was
disappointing because the car of UCD was unable to drive at the start of the
competition due to several issues. Fortunately the car drove the last day of
the competition. After year 3 the car must be in showroom condition. After
hard work the car was running. Unfortunately there was not enough time left
to finish the interior of the car. But the main goal is achieved: Better fuel
economy and improved performance. Year 4’s competition will be a long drive.
No decision is made yet where this event will take place.
UCD tries to go a step further than the goals of Challenge X: They like to built
a vehicle which only uses renewable energy. One of the events in Challenge X
is the ’Well to wheel’ event and this is UCD’s most important event.

3.1 Well to wheel petrolium use

The ’Well to wheel’ event looks at the total energy use of a car. Losses in every
step to get the energy to the the wheel will be taken into account. Energy
use will be calculated as an equivalent of petroleum usage. There are several
opportunities to get energy from different sources to the wheels. The idea of

9
 Figure 3.1: Stock Chevrolet Equinox 2005

UCD Davis is to transfer this energy as efficient as possible and maybe more
important: make use of renewable sources. The idea behind this plug in hybrid
is based on the use of solar panels: People can install a solar panel on their
roof. In one day you can collect enough solar power to charge the battery of the
car during the night. The next day the car can drive 40 miles all electric. This
is 15 miles more than the average American person drives per day. So, most of
the days you can drive your car for free without emissions. Of course the solar
panel is not free, but in five years you will earn the money back because you
don’t have to buy gasoline. Since an average solar panel will last 30 years, the
next 25 years your energy is free. When the batteries are discharged to about 20
percent SOC the car will drive in hybrid mode. The engine will always operate
on the ideal operating line and the electric motors will assist when necessary.
This driving mode results in a much more efficient fuel use than a conventional
car. For further decreasing emissions E85 is used as fuel. This year the team
was more successful than last year and especially in the ’Well to wheel’ event.
The team scored second on this event, with even more room for improvement,
which will be discussed later.

Figure 3.2: ”Plug in to the sun”

10
Chapter 4

Trinity

Trinity, the re-engineered vehicle, is a pre-transmission parallel plug-in hybrid.

The heart of the power train is a custom built Continuous Variable Transmission
designed at UC Davis. This CVT will be extensively discussed in 4.3.3. The
car is four-wheel driven: The front axle is driven by a Prius combustion engine
and a 75kW permanent magnet electric motor. Both are connected to the
input shaft of the CVT. The rear axle is driven by a 60kW induction electric
motor. Electric energy is stored in a 15.6 kWh lithium-ion battery pack. This
powertrain makes it possible that the vehicle performance and fuel economy of
Trinity is superior to the stock vehicle. Figure 4.1 shows a drawing with all the
major powertrain components.

Figure 4.1: Powertrain used in Trinity

11
4.1 Driving modes

Trinity has several driving modes: It can operate all electric (charge depleting),
charge sustaining and the so-called tow and four wheel mode. First of all the
car is designed in a way that an average American can drive an ordinary day
without burning oil: There is an electric range of 40+ miles all electric. When
the batteries reach the state of charge of 20 percent the car switches into hybrid
mode. The car also switches to hybrid mode when the speed exceeds 60 mph,
because the engine then operates in its most efficient spot. In case of all electric
mode the car is powered by the two electric motors. Together they can provide
135 kw. The front electric motor is used till it reaches its power limit, then the
rear electric motor will assist to fulfil the power command. The reason for using
the total power of the front electric motor first is the higher efficiency. This
higher efficiency is a result of transmitting the torque through the cvt. The ideal
operating line of the front electric motor will be used. Charge sustaining mode
starts when the batteries have a state of charge of 20 percent. This will prevent
the batteries from permanent damage. The engine is now the most important
power source. The electric motors will only assist in case of high accelerations
and for low speeds when the engine will shut down. To keep the state of charge
around 20 percent the engine will deliver a little more power than necessary
to drive the vehicle. A negative torque will be applied to the front electric
motor to charge the batteries. When the batteries reach a given maximum
state of charge (around 22 percent) the negative torque will be stopped. This
way the energy flow through the batteries is minimized. The batteries can be
fully charged by the grid, which is far more efficient. In the next section the
efficiency losses of the battery will be pointed out.

4.2 Electric motor and battery efficiency

As mentioned before the energy flow through the batteries should be minimized.
In every energy flow there are losses. The figures 4.2 and 4.4 in this section
show the difference between the motor efficiency with and without including
efficiency losses caused by the batteries. Figure 4.2 shows the motormap of a
Unique Mobility 100kW motor. This motor is not the same motor as used in
Trinity but a bigger version. For simulation this motor can be downsized but
that’s not necessary to show efficiency losses caused by batteries.
Motor efficiency is calculated as follows:

Tω
η= (4.1)
Eb Ib

Motor efficiency including battery losses is calculated as follows:

Tω Tω
η= = (4.2)
Eoc Ib (Eb + Ib Rb )Ib

12
 Figure 4.2: Efficiency map electric motor

In the first situation the mechanical output power is divided by the battery
voltage times the amount of amps. In the second situation the denominator
is represented by the open circuit voltage times the amount of amps. This
voltage is higher and will consequently result in a lower efficiency. In figure 4.3
a diagram is drawn which shows the difference between the two voltages.

Figure 4.3: Electric diagram battery and motor

The battery resistance in trinity is approximately 0.1 ohm. Losses due to this
resistance result in an overall lower efficiency for the motor. A new motormap,
including battery losses, is made. Figure 4.4 clearly shows the difference in
efficiency between either taking battery losses into account or not.

13
Figure 4.4: Efficiency map electric motor including battery losses

14
4.3 CVT

The custom built CVT by UCD is different from the conventional CVT on
several points. The used CVT in Trinity is a modified 2.0 L Jatco Continuously
Variable Transmission. The two main differences are the use of a chain instead
of a pushbelt and pulley actuation is done by servo pumps instead of an engine
driven pump. Next to these two differences the controls are done different too:
instead of controlling R, Rdot is controlled. These three differences will be
discussed in the next sections.

4.3.1 CVT chain

A conventional CVT transmits torque from the primary pulley to the secondary
pulley through a pushbelt. This pushbelt will slip what decreases the efficiency.
In the CVT of UCD the belt is replaced by a chain. This chain has a high
power and torque capacity, high efficiency (up to 94.7 percent) and produces
less noise. The chain is designed in a way that a linear motion of the chain itself
will result in a curved motion on the surfaces on the sides of the chain where it
interacts with the pulleys. A rolling action between the pins of the chain takes
place without slipping with respect to the pulleys, what avoids wear [Bro]. The
chain is produced by a company from the Netherlands: Gear Chain Industries.
Figure 4.5 shows the typical shape of the chain.

Figure 4.5: CVT chain by Gear Chain Industries

4.3.2 CVT hydraulics

The hydraulic system of the CVT is totally changed. In a conventional CVT

an engine driven pump delivers the required pressure to clamp the belt and
change the ratio. Electronically controlled valves controle the pressure for both
pulleys. At high speeds the pressure developed exceeds the required pressure,
so the pressure will be released. This pressure release contributes to losses. The
modified CVT uses high voltage servo hydraulic pumps for clamping pressure,
pressure for drive ratio and lubrication. Although there is still a little leakage,

15
losses are negligible compared to the conventional CVT. Calibration in this sys-
tem is easy; There is a pressure feedback form pressure transducers. Figure 4.6
shows the hydraulic diagram. The clamping pressure pump provides pressure
to both pulleys. The shift pump only provides pressure to change the driving
ratio.

Figure 4.6: Hydraulic diagram cvt

When the CVT case was totally machined by Team Fate members the hydraulic
lines had to be fabricated. First of all pipes from the inputs in the CVT (primary
pulley, secondary pulley and lubrication) to the NPT fittings in the case were
fabricated. With a pipe bender stainless steel pipes were bent till they fitted in
the small available space. The pipes are mainly located in the sump. All the
separate lines were pressurized to see if there were no leakages. This was done
with a drill which spinned a pump. All of them could withstand the required
pressure of 800 psi so the electric motor and the engine could be mounted
to the CVT, which was already mounted on the subchassis. When the front
powertrain was installed in the car the fabrication of hard -and softlines between
the CVT and the pumps could be started. Hard lines were used because it looks
better under the car. Important was to use a little piece of softline because the
pumps and the CVT can move with respect to each other; The pumps are rigidly

16
mounted to the subchassis and the cvt can move with respect to the subchassis.
The hardlines were made in the same way as the lines in the transmission itself.
Figure 4.7 shows the pipes installed under the car.

Figure 4.7: Pipes and hoses for CVT

When everything was properly installed and tested a cover of sheet metal was
made. This cover protects the powertrain from rocks and provides better aro-
dynamica.

4.3.3 CVT Control

The powertrain and its controls determine how the vehicle responds to the
driver’s commands. In Trinity the CVT is controlled different from the stock
CVT: By controlling the ratio rate of change, Rdot. The equation which de-
scribes the dynamic behaviour of the CVT is as follows:

−ṘIe ωe + RTe − Tdrag − Tlosses

ω̇ds = (4.3)
Ie R2 + Ids

The derivation of this equation can be found in appendix B

This equation is implemented in Matlab/Simulink. After implementing this
equation the model can be extended with a step in speed or a driver’s cycle and
a part of the used control strategy of Trinity. For better understanding of the
model a brief explanation of the control strategy is needed. The highest level of
control strategy shows that the car will first deplete the batteries till a state of
charge of 20 percent is reached. The engine will only fire up if the speed exceeds
60 mph, below this speed the electric motors supply the power. When the SOC
of 20 percent is reached the car will go into hybrid mode. This means that the
engine will be the major powersource, however the acceleration from 0 to 15
mph is done by the electric motors. When the car exceeds 15 mph the engine

17
starts and will operate on the Ideal Operating Line. The electric motors will
assist when necessary. In other words it will compensate for the lack of power
for hard accelerations. Furthermore the front electric motor delivers energy to
the batteries while breaking; regenerative breaking. This is not possible at the
rear axle due to the use of overrunning clutches. These are built in to avoid
drag of the electric motor while not being used. The SOC will be kept around
20 percent. The reason for this is already mentioned. As said, the dynamic
equation is implemented in a simulink model. The inputs for this equation are
R, Rdot and the engine torque. The other parameters of the equation will be
calculated or are constants. Output of the equation is the angular acceleration
of the drive shaft. From this acceleration the engine speed and vehicle speed
can be calculated. This will be the feedback signal of the model. A desired
vehicle speed is commanded. By subtraction the actual vehicle speed from the
desired vehicle speed an error is obtained. This error signal will be multiplied
by a gain. The output value of this multiplication represents Rdot. Rdot will
change the input of the dynamic equation what will affect the drive shaft speed.
The subsystem ’saturation’ avoids that the boundary values of R and Rdot will
be exceeded. The figures 4.8 and 4.9 show the model.

Figure 4.8: CVT model, commanding speed

The reason for using feedforward control is to avoid a short drop in speed when
the driver wants to accelerate. This drop in speed is a consequence of the term
with Rdot. The negative sign before the Rdot term makes this term exactly
the opposite of what is desirable. A CVT which is only powered by an internal
combustion engine requires an increase in engine torque. In a hybrid vehicle
this drop in speed can be compensated by an electric motor. This model and a
more extended model will show how to implement this.
A simulation with this model is done: A step in speed was commanded. Figure
4.10 will show the result of a step in speed from 10 to 15 m/s.
The results show that in a reasonable time the speed of 15 m/s is achieved. Rdot

18
 Figure 4.9: Saturation of R and Rdot

Figure 4.10: Results CVT simulation, commanding speed

is positive at the moment the speed command increases. This is what normally
happens in a CVT when an acceleration takes place. For this simulation a
constant engine torque was used. In the next model this will change.
The next step is to change into a model that is power commanded instead of
speed commanded. The power commanded model is shown in figure 4.11 and
figure 4.12. The first figure represents block ’CVT’ from the second figure.
Determining the desired power can be easily done by solving the equation for
Trl . In the simulink model this is done in the block ’Road Load’.

19
 Trl = Rw (fr mv g + 0.5rhoVv2 Cw A + mv a) (4.4)

This torque multiplied by the angular velocity of the wheels gives the desired
power. This signal will be commanded to the closed loop system, which is
situated in the CVT block. This block contains a subsystem with the CVT
dynamics. These dynamics are the same as the first model. Also the saturation
subsystem is the same. Output signal of the dynamic model of the CVT is
again the angular velocity of the drive shaft. To obtain the input shaft angular
velocity of the engine and motor angular velocity, driveshaft velocity will be
multiplied by the ratio. In a look-up table a torque belonging to the engine
or motor speed will be found. The feedback signal (the output power of the
engine plus the power of the motor) will be provided by the multiplication of
the engine and motor speed and the governing torque from the look up table.
The model uses two look up tables: One for the engine and one for the electric
motor. For lower speeds the electric motor is used and when the speed exceeds
15 mph the engine is started. Both look up tables should represent the Ideal
Operating Lines. Because the IOL of the electric motor is not available at this
moment a constant torque of 100 Nm is used. This can be easily changed in the
model when this information is available. Because the engine stays on the IOL
there will be a lack of power when a hard acceleration is required. The electric
motor will compensate for this lack of power: The power error in the model
is divided by the engine speed. This torque is added to the input shaft. Also
the compensation for the Rdot term will be a torque input on the input shaft.
These two compensations together is the electric motor torque, if the speed is
above 15 mph. Otherwise the torque obtained from the look up table of the
electric motor should be added too.

Figure 4.11: CVT control, power commanded

Running this model can show how the model follows a drive cycle. For the
simulation the USA drive cycle FTP-Highway is used. In figure 4.13 the results
will be shown of the power commanded CVT model.

20
 Figure 4.12: CVT model with drive cycle

Figure 4.13: Results CVT simulation, commanding power

It’s clear that the actual car speed is close to the desired speed. Only at a
couple spots a significant error can be noticed. Especially at the end a big error
can be seen but this can be easily explained. Braking in this model is only
done by the electric motor. When this motor reaches its maximum negative
torque there will be a lack of breaking torque. In the figure can be seen that
the electric motor reaches its maximum of -240 Nm a couple of times. This
happens at the end too. In a real car additional torque will be applied by the
disc brakes. Of course it’s possible to implement this in the model but that’s
not done yet. Another thing that has to be explained is the fact that R goes
fast to overdrive. This explanation is also an easy one. Driving the engine on
the IOL means that a high torque is delivered. The power required to drive the

21
car at a constant speed isn’t that high. This means that the engine speed has
to be low not to exceed the power request. This results in a low value of the
gear ratio (overdrive). At the start of the driving cycle only the electric motor
is working. When the car exceeds a certain speed the engine fires up and the
torque of the electric motor decreases. This control strategy makes the engine
run much more efficient than it does in a conventional car. So, IOL tracking
and the all electric range makes this car much more fuel efficient than the stock
vehicle. The model shown only represents the hybrid mode of the car.

22
Chapter 5

Prius engine

Where conventional cars get all their power form their internal combustion
engine, the engine in a hybrid electric car is assisted by an electric motor. This
fact makes it possible to use a smaller engine in such a vehicle. The goal of
building this car, as mentioned before, is increasing fuel economy, reducing
emissions and maintain or increase performance. The engine used is chosen due
to a couple of reasons: High thermal efficiency, relatively low emissions and its
compatibility with Ethanol (E85). The prius engine differs from a conventional
engine at the thermodynamic cycle: It runs on an Atkinson cycle. This cycle
provides higher thermal efficiency, lower emissions with the disadvantage of
losing power. In a hybrid configuration this can be compensated by an electric
motor.
To run an engine on the Atkinson cycle the camshafts have to be adjusted. This
has to be done in such a way that the intake valve is opened longer. This allows
intake air to flow back into the intake manifold. The effective compression
ratio will be reduced, so the expansion ratio will exceed the compression ratio.
A high expansion ratio allows a long powerstroke, so there is more time for
expansion of the combustion gasses and less heat is wasted. The peak thermal
efficiency increases from an average of 25 for the conventional engine to 37 for
the Atkinson engine.
When opening the cam cover an obvious difference can be noticed between the
cams for the intake valves and for the outlet valves. The cams for the intake
valves are less ’sharp’.

5.1 Why using ethanol

Nowadays people concern more and more about green house gas emissions. To
measure the amount of green house gasses emitted the following index is used:

GHGI = CO2 + 21 ∗ CH4 + 310 ∗ N2 O (5.1)

23
’The Greenhouse Gasses, Regulated Emissions, and Energy use in Transporta-
tion (GREET 1.5a) model was used to verify that the use of E85 in Trinity
would reduce greenhouse gas emissions by 43 percent compared to a conven-
tional vehicle fueled by gasoline [Sha07]. CH4 and N2O can be controlled with
after treatment, using a catalyst. Unfortunately this will not work for CO2, but
the use of ethanol instead of gasoline will reduce the CO2 emission considerably.
Another advantage of using ethanol is the higher octane number rating. Because
of the better knock resistant ethanol compression ratio can be increased, so
output power increases too.
Using ethanol means that the gas milage will be lower due to differences in
lower heating values. The lower heating value of gasoline is 31.5 MJ/L, where
the lower heating value of E85 is 22.6 MJ/L. This is 28 percent lower. To
inject the right amount of E85 the air-fuel ratios of gasoline and E85 have to
be compared. The air-fuel ratio of gasoline is 14.7, the air-fuel ratio of E85 is
10. This means that for every liter gasoline 1.47 liter E85 has to be injected to
burn a stoichiometric mixture. Injecting this amount of E85 will increase the
engine power with approximately 5 percent.

5.2 Adjustments on the engine

Some minor adjustments on the engine had to be made to make the engine
and its fuel system compatible with E85. Ethanol is highly corrosive. A new
E85 compatible fuel tank was bought. All the fuel lines are made of stainless
steel and the fuel pump is a special pump for alcohols which can provide a
constant pressure of 60 psi. This pressure is necessary to vaporize the ethanol.
Furthermore the injector O-rings wre replaced by neoprene rubber O-rings.
These adjustments were the only advices from the dealership. Other modifica-
tions made on the engine will be presented later, because these have nothing to
do with the fuel change.

24
Chapter 6

Dynamometer test stand

The hybrid electric vehicle lab already had a testrig with an eddycurrent dy-
namometer but a more heavy duty dynamometer was desirable. Two major
projects have to be done in the near future with this dynamometer: Tuning a
Prius engine for ethanol combustion and test a pre-transmission parallel hybrid
drivetrain with an inline CVT. The goal of the second project is to test the
controls for this drivetrain. At this moment the team is integrating the same
drivetrain in Sequoia, another car of Team Fate. This project requires a more
heavy duty dynamometer, while the Prius engine has a maximum power of 59
kw and a maximum torque of 111 Nm. For the permanent parts of the dy-
namometer test stand everything had to be designed for the maximum torque
the dynamometer can handle.
In the first place the dynamometer will be used for tuning the prius engine.
During competition the car was driving well but the engine was not tuned yet
for E85. The power of this hybrid design is to run the engine constantly on
its ideal operating line. The purpose of this test stand is to obtain 2 figures:
Throttle position to torque and an engine map. This will provide the necessary
information to control the drivetrain in such a way that the engine will always
run on the IOL.
To get the test stand operational a lot of things had to be done. These things
are described in the next paragraphs.

6.1 Brackets and shafts

Several brackets and shafts had to be made. The permanent parts had to be
designed for the maximum torque the dynamometer can handle. A decision
was made to couple the dynamometer and the engine with a double U-joint.
This joint can compensate for misalignment in horizontal and vertical direction.
For the project with the CVT drivetrain there was already a male spline shaft
with a flange which was made to connect to the U-joint. This mail spline shaft
connects to the output shaft of the CVT. This part could be used for the Prius

25
engine project too: A female spline with a flange was made. This flange can
connect to the flywheel. Now only a connection between de dynamometer and
the U-joint had to be made. In the next paragraphs each part made will be
discussed. But first some analyses were done on the shafts which were already
in the lab.
The U-joint can compensate for misalignment of several millimeters. The di-
mensions of this shaft are shown in the next picture.

Figure 6.1: Dimension of the double U-joint

In report [Sch05] is written about the male spline shaft. The spline, shown in
figure 6.2 is made of steel 4340 and has a key connection on the other side.
Two calculations show that this shaft and key connection are strong enough to
withstand the maximum torque which can be applied to the dynamometer.

Figure 6.2: The spline shaft with key connection

26
The governing equation for maximum torque applied on the shaft before failure
occurs is as follows:
√
πd3 3
Tmax,shaf t = Sy (6.1)
16 3

Solving this equation with the right dimensions and the right material properties
of the shaft gives a maximum torque of 1063 Nm.
The governing equation for the maximum allowable torque on the key is as
follows:
√
3 2
3 Sy Ld
Tmax,key = (6.2)
8

Solving this equation results in a maximum torque of 1354 Nm. Both values
are higher than the maximum torque the dynamometer can load to the system:
948 Nm. In fact this torque will never be reached because the first project can
never deliver this amount of torque and the second project’s purpose is to verify
the controls. Such a high torque is not necessary for this purpose.
After this analyses the brackets and other shafts had to be made.

6.1.1 Engine mounting brackets

The engine had to be rigidly mounted to the stand. Only one stock mounting
point was on the back of the engine. This is one of the points where the engine
will be mounted to the stand. But first the front of the engine will be mounted
to a bracket. This bracket had to be designed to withstand the most torque
the engine produces. Out of a big 1/4” steel plate from the scrap pile a circle
with some kind of wings at two sides was cut with the cutting torch. In the
circle a hole was cut with the diameter of the flywheel plus a little clearance.
On the circle 8 holes were drilled to mount to the engine. To locate the centers
of the holes a drawing with the exact locations of the holes was plotted and
attached to the material. This drawing is shown in appendix D. Under the
two wings little plates with two holes were welded. These plates were mounted
with 2 bolts to already existing brackets, consisting rubber damping mounts.
See figure 6.3.
Considering the maximum torque of 111 Nm the maximum stress in the bolts
of the bracket can be calculated. Information about maximum tensile strength
of bolts was found at www.mcmaster.com. The used bolts have a minimum
tensile strength of 120.00 [psi], 827.4 [MPa]. In this case the shear stress is
the stress to be analyzed. The maximum tolerable force can be calculated as
following [Gro05]:

Fmax = 0.7σ2A (6.3)

27
 Figure 6.3: Front engine mounting bracket

π 2
A= D (6.4)
4

Where A is 7.85e-5 m2 , the maximum force on one bolt is 90 kN.

The distance from the shaft to the bolts is 168 mm. At the maximum torque the
force on the radius of the bolts is just 655 N. This will not cause any problem.
Because of the chosen thickness of the steel there will be no problem at all with
stresses in this bracket.
At the rear of the engine another bracket was made. The most important goal
for this bracket is to just carry the engine: Adding a vertical force. Because
there also will be a little bit of torsion, bars were mounted on 45 degrees to the
main pilar. All the parts are made of steel, rectangular profile. The fabrication
of this part was done when the engine was already mounted to the front mount.
The rear was carried by a lift. The bracket is shown figure 6.4.

28
Figure 6.4: Rear engine mounting bracket

29
6.1.2 Dynamometer mounting brackets

Maximum torque applied on the dynamometer could be 700 foot pound, what is
equal to 948 Nm. With this knowledge brackets could be designed. A decision
was made to make 3 separate brackets. Two of these are the same and were
mounted to a quarter inch steel plate which was mounted to the blue disk on
the dynamometer. See figure 6.5. Fortunately there were two old steel brackets
available which could be used after some little grinding work. The other one
is mounted under the dynamometer. This one has to carry a part of the own
weight of the dynamometer. The other 2 brackets are most important. Together
they must withstand the torque applied on the dynamometer. The quarter inch
plate was mounted with 4 bolts to the blue disk. To be sure the brackets, the
bolts and especially the metal around the bolts are strong enough some stress
analyses had to be done.
The radius where the torque applies on the bolts on the steel plate is 0.12 [m]
for the connection to the dyno. When the total force will apply on just one
bolt the resulting force on that bolt will be 7.9 [kN]. The maximum force can
be calculated similar as in the previous paragraph. This time A = 5.1e-4, what
results in a maximum force of 590 kN. These bolts will not give a problem, so
the bolts on the brackets which are further from the axis will not give problems
either.
The brackets where the plate is mounted at are just 3.5 mm thick. The area
where the force of the bolts is working at is determined as follows:

1
A= ∗ t ∗ circumf erence (6.5)
3

Here t is the thickness of the material of the bracket. A=4.4e-5. Stress is the
force divided by the surface. On this radius the force is 3.8 kN. This will result
in a stress of 86 MPa. This stress will be reached in case of only one loaded
bolt. The yield stress of steel is 400 [MPa] so this may not cause any problems.
A picture of the dynamometer mounting is shown below.
The bracket to carry the own weight of the dynamometer is also made of steel.

30
Figure 6.5: Mounting dynamometer

31
6.1.3 Dynamometer adaptor plate

Next to the brackets the connections between the flywheel and the dynamometer
had to be made. First an adaptor was made to connect the dynamometer to
the U-joint. At the dynamometer side it connects to a surface with 8 threaded
holes. At the U-joint side it has to connect to a surface with 4 through holes.
The first design consisted a shaft with a flange on both sides. There were two
ways to machine this part: in one part or 3 separate parts. Machining it from
one part on the lathe will result in wasting a lot of material and will take a long
time. First machining 3 separate parts and weld them together was the other
possibility. Probably this is the cheapest and the fastest way but concentricity
will be an issue. Welding the parts exactly concentric will be almost impossible.
A new design was made: Two separate disks which should bolt together. One
piece has the 8 holes for the connection to the dynamometer and the other
one the 4 holes for the connection to the U-joint. Both pieces are one inch
thick. In both pieces a circular pocket of half an inch was made. This is the
area where the bolts will come. On a bigger radius 4 holes were made in both
pieces. These holes make it possible to bolt the pieces together. The heads of
the bolts which connect to the dynamometer or ujoint are inside the piece. To
obtain perfect concentricity connecting the parts a little step in thickness was
made. The first piece has a thickness of, from outer to inner diameter, 1 inch
to 0.9 inch to 0.5 inch. The other piece will will have a thickness of, from outer
to inner diameter, 1 inch, 1.1 inch and 0,5 inch. A potential misalignment is
prevented. All the holes are made on the mill. With a special indicator the
center of both parts could be located. A little program was written to drill the
holes at the right positions. Drawings of both parts are shown in 6.6 and 6.7.
A picture of the parts bolted down to the dynamometer and the U-joint can be
seen in the previous paragraph.

Figure 6.6: Dynamometer adaptor part 1

32
 Figure 6.7: Dynamometer adaptor part 2

Each bolt can withstand a force of 147 kN. Dividing the maximum force by
the radius shows that the maximum force will be 22 kN so the bolts are strong
enough.

6.1.4 Flywheel adaptor shaft

Only one mechanical part lasts now. The connection between the flywheel and
a male spline. The flywheel is bolted to the crankshaft with 6 bolts. This is
the place where the shaft has to connect to. A flange of 3/8 inch thick will be
bolted down to the flywheel. The other end consists of a female spline, made by
Gear Industries. Only the 6 holes of this part are made in the Student Machine
Shop. These holes were also drilled on the mill by writing a little program. A
drawing and a picture are shown in the figures 6.8 and 6.9. In the drawing the
spline is not shown, but the manufacturer used the male spline to make this
one.
This shaft completes the connection between the engine and the dynamometer.
Unfortunately the bolts to connect the shaft to the flywheel didn’t arrive before
leaving the lab, so a picture with everything bolted together could not be made.

33
 Figure 6.8: Flywheel adaptor shaft

Figure 6.9: Picture of the flywheel adaptor shaft connected to the splineshaft
for the cvt

34
6.2 Engine

6.2.1 Engine controls

In Trinity the engine is controlled by a Mototron engine controller with 128

pins. This controller was selected because of the robust automotive production
and the possibility to produce the control algorithms with Simulink. These
controls can run on hardware with a real time operating system. This allows
students to develop their controls in a simulation environment. One of the tools
of MotoHawk will automatically generate a code from Simulink models which
the hardware controller uses. The engine controller in Trinity takes care of
all the necessary engine controls and some accessories like cooling pumps. In
trinity an electric pump is used for the engine cooling but for the dynostand
the stock waterpump will be used. The mototron controller makes it possible
to control the engine for ethanol combustion.
As said, the engine controller takes care of all the engine controls, which means
reading several sensors and actuating several systems. In appendix C the con-
troller is shown with all its in and output ports and the wiring to the sensors,
coils and fuelinjectors. For the dynostand a total new wiring harness was made
to control this engine. The wire diagram was already available from the car,
only some minor changes were made. The controller was placed close to the
engine. From here necessary wires go to the engine, battery and to the out-
side of the dynocell. The wiring going outside of the dynocell are the wires for
the keyswitch, CAN (communication area network), serial, fuelpump and the
startermotor. This allows to start the engine without being in the dynocell.
When the engine is running nobody is supposed to be there.
Due to the use of this standard engine controller some adaptions had to be made
on the engine. The Mototron controller is adjustable, but at some levels in the
software it’s impossible to change the standard controller. Because of this a new
encoder plate for the crankshaft position sensor was installed. Mototron only
support a couple of these encoder plates, so a right encoder plate was ordered
from Saab. The encoder plate was mounted to the crankshaft pulley with an
adaptor plate which was made on the lathe. The material used is aluminum,
so it doesn’t weight too much. A picture of this adaptor plate and the new
encoder plate is shown in figure 6.10. By making a little step in height with
the diameter of the hole in the encoder plate concentricity was ensured. The
encoder plate and the adaptor were bolted down by 4 bolts and this together
was bolted down by the bolt of the crankshaft. This allowed to change the
angular positioning of the encoder plate by releasing the crankshaft bolt.
For the same reason as the encoder plate the encoder for the camshaft position
had to change. By taking of the camcover 3 encoder pins can be seen. This
had to change into one pin. The middle one was the necessary one, the other
two were pressed out. Figure 6.11 shows the original camshaft where are the
pins are still there.

35
Figure 6.10: Encoder plate withe the adaptor plate

Figure 6.11: Encoder pins on camshaft

36
6.2.2 Engine cooling

Engine cooling is done by the stock water pump. Where usually a radiator is
used to cool the engine coolant, now a heat dump will be used. This heatdump
consists of a reservoir, a waterinlet from the sink, a wateroutlet to the drain
and an inlet and outlet for the water circulation through the engine. The water
temperature in the reservoir is measured and when the temperature gets to
high water from the reservoir goes into the drain and cold water from the sink
will flow into the reservoir. The advantage of this system is that the small
dynocell will not be heated. In this case it’s actually necessary because the
cooling facilities in the small room are not sufficient.
The heatdump is shown in figure 6.12.

Figure 6.12: Heatdump

6.3 Dynamometer

The hybrid electric vehicle lab obtained an old ’Go Power Systems’ waterbrake
dynamometer. With this waterbrake a solenoid valve and a display/control
console came. Unfortunately this console was so old that is was impossible to
order the products to get it operational. The specifications are comparable to

37
a newer version from Go Power Systems: The 516 Series.

Torque 700 LB-Ft or 948 Nm

Power 900 HP
RPM 7500 Peak and 5000 Cont.
Rotation CW, CCW
Alloy Material Aluminum
Water Requirements 3/4” - 1.0 line, 35 psi, .06 GPM/HP
Weight 95 LB
Torque Transducer Hydraulic
The control systems Go Power sells are very expensive so an own controller
had to be made and used. Because the dynamometer hadn’t been used for
a long time a load cell refill kit and new load cell diaphragms were bought
and installed. It was decided not to buy a new display console because the
measurements can be displayed on a pc as well.

6.3.1 Hydraulic system of the dynamometer

The last part to get this system operational is building the hydraulic systems
for the dynamometer and control it. There are two seperate hydraulic system
necessary to get the water brake to run. The first system is important for data
acquisition. Next to a tachometer there are two load cells. These two cells
together apply pressure which represents the torque applied to the dynamome-
ter. A hose from both of the load cells come together in a tee fitting. At the
other end of the tee fitting a hose coupled to a pressure transducer is installed.
This pressure indicates the amount of torque on the waterbrake. The system
is designed so that one psi represents one lb-ft. The pressure transducer for
this system is ordered from the company Digikey. The specifications of this
transducer are in the table below.

Pressure range 1000 psi

Pressure unit psi
Electrical output 0.5 to 4.5 Vdc
Pressure port 1.4 inch NPT
Electrical connection 2ft cable
The pressure range is chosen so that it just exceeds the maximum pressure the
load cells can apply. The electrical output is chosen to give the right output for
the used Mototron controller. At the moment fluid flows through flexible hoses
but these will be substituted by stainless steel brake lines. These brake lines
will be flared and a NPT fitting will be attached so it fits the 1/4 inch NPT
fitting of the transducer.
Another hydraulic system has to take care of the right water flow into the dy-
namometer. The water level in the brake determines how much power can be
absorbed. To control this input flow a pump, valve and a reservoir are necessary.

38
Output flow control is not necessary. Considering the water requirements of the
dynamometer a pump was bought. Because all small commercial pumps meet-
ing this requirements were very expensive an oversized lawn pump is bought
from Wayne pumps. The specifications of this pump are shown in the table
below.

Continuous duty pump has 1 1/2 HP 230 Volt, single stage motor
Dual voltage available (115/230), factory wired for 230 volts
Motor Speed: 3450 RPM
Cast iron pump housing stainless steel shaft for durability
Maximum Suction Lift: 25ft.
Max PSI: 50 PSI
2in. NPT suction port
1 1/2in. NPT discharge port
Moves up to 4790 GPH @10 PSI 5ft. lift and 2390 GPH @ 30 PSI 25ft. lift
Glass reinforced thermoplastic impeller and diffuser for corrosion-resistance
Ceramic/carbon-faced seal for long life
Besides a pump a reservoir of 9 gallon was ordered. Unfortunately this part was
not delivered before leaving, so the connection between different components
with pipes and hose hasn’t been made yet. The first design of how the different
components have to connect to each other is shown in figure 6.13.

Figure 6.13: Dynamometer hydraulics

As mentioned before the cooling in the dyno cell is not sufficient. This is the
reason for not using a radiator, so the principle of the heatdump will be used

39
again. A temperature sensor will monitor the water temperature at the output
of the dynamometer and when this becomes to high cold water will flow into
the reservoir and push out warmer water, which goes directly to the drain. To
regulate the inlet flow a solenoid valve from the company Pulsecooling will be
used. This valve receives a constant 12VDC signal. Grounding one of the pins
will open the valve.

6.3.2 Controlling the dynamometer

As said the dynamometer absorbs an amount of power depending on the water

level in it. The water level changes as a consequence of difference between the
input and the output flow. This can be done by controlling the input flow.
The torque load of the waterbrake is correct when the rotational speed stays
constant. To control this inlet flow a solenoid valve will be used. A proportional
flow valve would be more ideally but the solenoid valve was delivered together
with the dynamometer, so it’s a cost saving issue. There will be a constant
pressure from the pump. The valve will be opened if necessary. Pulses will be
sent to the valve to achieve an as good as possible constant flow. Dependent
on the required flow the valve will be opened for a longer period. Changing the
length of the pulses will be done according to the rotational speed. Just like
with the engine, a Mototron controller will be used. This time it will be a 48
pin controller of Mototron. The reason for using a second controller is that the
waterbrake will be permanent and several projects will use this dynamometer.

6.4 Results

Unfortunately there was not enough time to finish this project. This means
there are no results to show, but a lot of work has already been done. The
engine is ready to start if all the wires which are already made will be connected
to the Mototron controller. The mechanical connection between the engine
and the dynamometer is completed and all the necessary brackets are made.
Completing the hydraulic system for the water supply is all that is left to do.
In the near future the test stand will be operational, so an engine map and a
throttle-torque curve can be produced.

40
Chapter 7

Conclusion

The report shows that a lot of work has been done but that the actual assign-
ment is not finished yet unfortunately. A lot of time and work went into the
competition and not without good results. The team did not end up high at
the ranks, but that was not the main goal. A better fuel economy was achieved,
with even more room left to improve. For better understanding of the drive-
train a model of the CVT and its controls were made. The drivecycle can be
followed well, which shows that this control strategy can really work. Results
clearly show how the internal combustion engine and front electric motor are
working together. The model can be extended to calculate fuel economy for
different drive cycles. Although the test stand is not finished yet, most of the
work has been done. Brackets and shafts were made, all parts were ordered and
the engine is almost ready to be fired up. When the total hydraulic system of
the dynamometer is hooked up, controls have to be written and then the stand
is ready to use.

41
42
Appendix A

Symbols

η efficiency -
T Torque Nm
ω angular velocity rad/s
Eb battery voltage V
Ib battery current amps
Eoc open circuit voltage V
Ioc open circuit current amps
Rb battery resistance ohm
ωds angular velocity drive shaft rad/s
R cvt ratio -
Te cvt input torque Nm
Ie engine inertia kgm2
Ids drive shaft inertia kgm2
Trl torque road load Nm
Rw wheel radius m
fr rolling resistance coefficient -
mv vehicle mass kg
g gravitational acceleration m/s2
ρ mass density kg/m3
Vv2 vehicle speed m/s
Cw air drag coefficient -
A frontal area vehicle m2
a vehicle acceleration m/s2
Fmax maximum allowable force N
d diameter m
Sy yield stress Pa
σ stress Pa
L length m
t thickness m

43
44
Appendix B

Derivation of the acceleration

of the drive shaft

Figure B.1: Schematic CVT drive train

Ie ω̇e = Te − Tp (B.1)

Ids ω̇ds = Td − Tds (B.2)

Td − Tds Tp r − Tds
ω̇ds = = (B.3)
Ids Ids
(Te − Ie ω̇e )r − Tds
ω̇ds = (B.4)
Ids
ωe = rωds (B.5)

ω̇e = ω̇ds r + ωds ṙ (B.6)

(Te − Ie (ω̇ds r + ωds ṙ))r − Tds
ω̇ds = (B.7)
Ids
(Ids + Ie r2 )ω̇ds = Te r − Ie rωds ṙ − Tds (B.8)

45
 Te r − Ie rωds ṙ − Tds
ω̇ds = (B.9)
(Ids + Ie r2 )
Te r − Ie ωe ṙ − Tds
ω̇ds = (B.10)
(Ids + Ie r2 )

46
Appendix C

Engine harness

Figure C.1: Engine harness

*Some of the wires are not shown. This to keep the drawing readable. Color

47
and pin on the Mototron are mentioned. Engine can fire up without a knock
and oil pressure sensor but in the feature these will be added.

48
Appendix D

Technical drawings

D.1 Dynamometer adaptor plate part 1

49
Figure D.1: Technical drawing dynamometer adaptor plate part 1

50
D.2 Dynamometer adaptor plate part 2

Figure D.2: Technical drawing dynamometer adaptor plate part 2

51
D.3 Flywheel adaptor shaft

Figure D.3: Technical drawing flywheel adaptor shaft

52
D.4 Encoder adaptor plate

Figure D.4: Technical drawing encoder adaptor plate

53
D.5 Front engine

Figure D.5: Technical drawing holes front engine

54
List of Figures

3.1 Stock Chevrolet Equinox 2005 . . . . . . . . . . . . . . . . . . . . 10

3.2 ”Plug in to the sun” . . . . . . . . . . . . . . . . . . . . . . . . . 10

4.1 Powertrain used in Trinity . . . . . . . . . . . . . . . . . . . . . . 11

4.2 Efficiency map electric motor . . . . . . . . . . . . . . . . . . . . 13
4.3 Electric diagram battery and motor . . . . . . . . . . . . . . . . 13
4.4 Efficiency map electric motor including battery losses . . . . . . 14
4.5 CVT chain by Gear Chain Industries . . . . . . . . . . . . . . . . 15
4.6 Hydraulic diagram cvt . . . . . . . . . . . . . . . . . . . . . . . . 16
4.7 Pipes and hoses for CVT . . . . . . . . . . . . . . . . . . . . . . 17
4.8 CVT model, commanding speed . . . . . . . . . . . . . . . . . . 18
4.9 Saturation of R and Rdot . . . . . . . . . . . . . . . . . . . . . . 19
4.10 Results CVT simulation, commanding speed . . . . . . . . . . . . 19
4.11 CVT control, power commanded . . . . . . . . . . . . . . . . . . 20
4.12 CVT model with drive cycle . . . . . . . . . . . . . . . . . . . . . 21
4.13 Results CVT simulation, commanding power . . . . . . . . . . . 21

6.1 Dimension of the double U-joint . . . . . . . . . . . . . . . . . . 26

6.2 The spline shaft with key connection . . . . . . . . . . . . . . . . 26
6.3 Front engine mounting bracket . . . . . . . . . . . . . . . . . . . 28
6.4 Rear engine mounting bracket . . . . . . . . . . . . . . . . . . . 29
6.5 Mounting dynamometer . . . . . . . . . . . . . . . . . . . . . . . 31
6.6 Dynamometer adaptor part 1 . . . . . . . . . . . . . . . . . . . . 32
6.7 Dynamometer adaptor part 2 . . . . . . . . . . . . . . . . . . . . 33
6.8 Flywheel adaptor shaft . . . . . . . . . . . . . . . . . . . . . . . . 34

55
6.9 Picture of the flywheel adaptor shaft connected to the splineshaft
for the cvt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
6.10 Encoder plate withe the adaptor plate . . . . . . . . . . . . . . . 36
6.11 Encoder pins on camshaft . . . . . . . . . . . . . . . . . . . . . . 36
6.12 Heatdump . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
6.13 Dynamometer hydraulics . . . . . . . . . . . . . . . . . . . . . . . 39

B.1 Schematic CVT drive train . . . . . . . . . . . . . . . . . . . . . 45

C.1 Engine harness . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

D.1 Technical drawing dynamometer adaptor plate part 1 . . . . . . 50

D.2 Technical drawing dynamometer adaptor plate part 2 . . . . . . 51
D.3 Technical drawing flywheel adaptor shaft . . . . . . . . . . . . . 52
D.4 Technical drawing encoder adaptor plate . . . . . . . . . . . . . . 53
D.5 Technical drawing holes front engine . . . . . . . . . . . . . . . . 54

56
Bibliography

[Bro] A.W. Brown, J.v. Rooij, and A.A. Frank. The design of an inline gci
chain cvt for large vehicles. Technical report, Brown Co., GCI and
UCDavis.

[Gro05] M.P. Groover. Fundementals of Modern Manufacturing Material,

Processes, and Systems, London: Prentice Hall International Editions.
2005.

[Sch05] Niels Scheffer and Guus J.C.M. Arts. Realisation inline-cvt testrig.
Internship report, UC Davis Hybrid Electric Vehicle Center, 2005.

[Sha07] Andrew Shabashevich, Douglas Saucedo, Terrence Williams, Christian

Reif, Cuyler Lattoraca, Bryan Jungers, Beth Wietzel, Professor An-
drew Frank, and Team Fate. Consumer ready plug-in hybrid electric
vehicle. Technical report from competition, University of California,
Davis, 2007.

57