The University of Toledo
The University of Toledo Digital Repository
Theses and Dissertations
2005
An ultracapacitor-battery energy storage system for
hybhrid electric vehicles
Adam W. Stienecker
The University of Toledo
Follow this and additional works at: http://utdr.utoledo.edu/theses-dissertations
Recommended Citation
Stienecker, Adam W., "An ultracapacitor-battery energy storage system for hybhrid electric vehicles" (2005). Theses and Dissertations.
1464.
http://utdr.utoledo.edu/theses-dissertations/1464
This Dissertation is brought to you for free and open access by The University of Toledo Digital Repository. It has been accepted for inclusion in Theses
and Dissertations by an authorized administrator of The University of Toledo Digital Repository. For more information, please see the repository's
About page.
A Dissertation
entitled
AN ULTRACAPACITOR - BATTERY ENERGY STORAGE
SYSTEM FOR HYBRID ELECTRIC VEHICLES
by
Adam Stienecker
Submitted as partial fulfillment of the requirements for
the Doctor of Philosophy Degree in Electrical Engineering
___________________________
Advisor: Dr. Thomas A. Stuart
___________________________
Graduate School
The University of Toledo
August 2005
An Abstract of
AN ULTRACAPACITOR - BATTERY ENERGY STORAGE
SYSTEM FOR HYBRID ELECTRIC VEHICLES
Adam W. Stienecker
Submitted as partial fulfillment of the requirements for
the Doctor of Philosophy Degree in Electrical Engineering
The University of Toledo
August 2005
The nickel metal hydride (NiMH) batteries used in most hybrid electric vehicles
(HEVs) provide satisfactory performance but are quite expensive. In spite of their lower
energy density, lead acid batteries are much more economical, but they are prone to
sulfation in HEV applications. However, sulfation can be greatly reduced by a circuit
that uses an ultracapacitor in conjunction with the battery. This research presents a new
cost-effective method for using these two energy storage components together in order to
extend the life of the battery. This system is presently quite expensive, but it will provide
much cheaper energy storage if ultracapacitor prices can be reduced to the levels
predicted by some manufacturers.
This dissertation studies two different methods for implementation on a hybrid
electric vehicle and presents performance data for a variety of simulations.
ii
ACKNOWLEDGEMENT
This research was supported by a research grant from DaimlerChrysler, AG and
by NASA Grant NAG3-2790 under subcontract from Bowling Green State University.
iii
Table of Contents
ABSTRACT........................................................................................................................II
ACKNOWLEDGEMENT ................................................................................................ III
TABLE OF CONTENTS.................................................................................................. IV
LIST OF FIGURES ........................................................................................................... V
CHAPTER 1 INTRODUCTION ...................................................................................... 1
CHAPTER 2 THE ULTRACAPACITOR (UC) .............................................................. 3
CHAPTER 3 THE BATTERY ......................................................................................... 5
3.1: Failure in Lead Acid Batteries ......................................................................................................7
CHAPTER 4 PRESENT UC – BATTERY HYBRID ENERGY STORAGE SYSTEMS
........................................................................................................................................... 10
CHAPTER 5 THE PROPOSED UC – BATTERY ENERGY STORAGE SYSTEM... 14
CHAPTER 6 BATTERY CHARGING METHODS FOR THE PROPOSED SYSTEM
........................................................................................................................................... 20
CHAPTER 7 PERFORMANCE CHARACTERIZATION METHODS ....................... 25
7.1 The Vehicle Simulation ..................................................................................................................27
7.2 HEV Control Strategy....................................................................................................................30
7.3 System Software .............................................................................................................................33
7.4 System Hardware...........................................................................................................................43
CHAPTER 8 EXPERIMENTAL RESULTS ................................................................. 47
CHAPTER 9 CONCLUSION......................................................................................... 59
REFERENCES: ................................................................................................................ 61
APPENDIX I: SYSTEM SOFTWARE ............................................. ON ENCLOSED CD
A.
B.
C.
D.
Primary Microcontroller Software..................................................................... On Enclosed CD
Secondary Microcontroller Software ................................................................. On Enclosed CD
PC Monitor Software ......................................................................................... On Enclosed CD
ABC-150 ROS Software ..................................................................................... On Enclosed CD
APPENDIX II: SYSTEM SCHEMATICS...................................................................... 69
iv
List of Figures
Figure 1.1
Hybrid Fuel Cell – Ultracapacitor ESS
2
Figure 1.2
Hybrid Battery-Ultracapacitor ESS
2
Figure 4.1
Equivalent circuit of a parallel battery/UC
11
Figure 4.2
Hybrid UC-Battery Simulated Discharge Current Waveforms
12
Figure 5.1
Proposed Hybrid Battery-UC ESS
15
Figure 5.2
Simulated Constant Power Pulse
18
Figure 6.1
First Battery Charging Method
20
Figure 6.2
Second Battery Charging Method
21
Figure 7.1
The Proposed Performance Characterization Method
26
Figure 7.2
ICE Efficiency Map and The Power Split Rule
31
Figure 7.3
Battery Charging and Routine Message Transmission Algorithm
34
Figure 7.4
Analog Interface and Vehicle Simulation Algorithm
35
Figure 7.5
Secondary Microcontroller Algorithms
39
Figure 7.6
PC Monitor Software screen shot
40
Figure 7.7
ABC-150 ROS Algorithm
43
Figure 7.8
Battery Voltage Sense signal conditioning
44
Figure 7.9
IR2118 Based Relay Driver Circuit
45
Figure 7.10
LEM Current Sensor signal manipulation circuit
46
Figure 8.1
Results for 300A / 5 sec. test currents
48
Figure 8.2
Results for 300A / 8 sec. test currents
49
Figure 8.3
Results for the Idle-Stop Test
50
Figure 8.4
StampPlot Pro screen capture
51
Figure 8.5
Simulator results to demonstrate battery charging and
ICE Only Mode
52
Figure 8.6
Simulator results to demonstrate Power Assist Mode
53
Figure 8.7
Simulator results to demonstrate Motor Only Mode
54
Figure 8.8
Simulator results to demonstrate regenerative braking
55
Figure 8.9
Buck Regulator Parallel Battery Charging System
56
Figure 8.10
Results of the Parallel Battery Charger
57
v
Chapter 1
INTRODUCTION
Due to increasing fuel costs and the rising governmental restrictions on pollutants,
automobiles have become increasingly more sophisticated and expensive. One of the
more recent advances in the automobile industry is the current push for hybrid power
within the vehicle. Many automobile manufacturers are attempting to compete in this
new arena by implementing an Energy Storage System (ESS) and an electric motor along
side the common-place Internal Combustion Engine (ICE). This approach is commonly
termed the Hybrid Electric Vehicle (HEV). During a typical driving cycle the HEV
efficiently uses energy from both the gasoline, through the ICE, and the ESS, through the
electric motor. A recent advance within the HEV sector is the hybridization of the
Energy Storage System [1].
The Hybrid Energy Storage System (HESS) consists of two or more different
forms of energy storage used to power the electric motor within the HEV. Given this
definition, limitless possibilities can be imagined. Throughout the world, research has
been conducted on these hybrid systems, such as a Hybrid Fuel Cell – Ultracapacitor
ESS[2,3] and the Hybrid Battery – Ultracapacitor ESS[4,5,6,7,8,9,10] to name a few. This
research has shown an improvement in the overall efficiencies of these
vehicles[2,3,4,5,6,7,8,9,10].
1
2
DC/DC
Converter
PEM
Fuel Cell
DC/DC
Converter
Power
Inverter
Hydrogen
Storage
Ultracapacitor
Battery
Electric
Motor
Electric
Motor
Figure 1.1– Hybrid Fuel Cell –
Ultracapacitor ESS[2]
Power
Inverter
Ultra+
+ Capacitor
Figure 1.2 – Hybrid Battery –
Ultracapacitor ESS
Due to different energy storage systems present within one HEV there is the
possibility of multiple voltage levels with the same electrical system. Some
implementations of the HESS have utilized DC-DC Converters[11] to create a single
voltage for the entire system and, therefore, a single voltage input to the motor drive. In
the simplest system architecture of the Hybrid Battery – Ultracapacitor ESS a single level
voltage is implemented. This is done by simply attaching the Ultracapacitor in parallel
with the Battery[4,8,9]. It is generally more cost effective if all energy storage devices are
kept at the same voltage level to avoid the need for complex voltage conversion and
switching devices. However, it now appears possible to develop a simpler and cheaper
scheme by using multiple voltage levels within one system.
Chapter 2
THE ULTRACAPACITOR (UC)
Due to advances in technology[2,12,13,14,15], manufacturers are now able to produce
capacitors in the 2700+ Farad range at a cell voltage of about 2.5 Volts. However, at
present energy densities of only 5-10 watt-hours per kilogram[13] the UC seems to be a
very poor choice, compared to the 50+ watt-hours per kilogram[9] available in a NickelMetal-Hydride (NiMH) battery, which is used in many HEVs.
On the basis of energy density alone the battery is clearly the better choice, but
when power density is considered along with energy density the choice is not so clear.
The main reason for the large difference in power density between the battery and the UC
is the method used to store the energy. In a battery energy is stored through a chemical
reaction. When a battery is charged the energy stimulates a chemical reaction which is
reversed when energy is removed from the battery. This chemical reaction is the main
reason that a battery has a lower power density. UCs, on the other hand, store energy in a
completely different manner. The UC uses charge separation as its method for storing
energy [12]. This allows the energy to be stored and released without any chemical
reaction taking place. This makes the UC a desirable energy storage device because it
can accept and release energy very fast and with low losses. UCs have been produced
with a maximum power density higher than 5 kW/kg as compared to a typical NiMH
3
4
battery with a maximum power density of 1/10 that of the UC or about 0.5 kW/kg, for a
matched impedance discharge [12].
The availability of UCs over the last 3-4 years has steadily improved. Several
manufacturers now produce UCs commercially [16,17,18,19,20,21]. Other manufacturers also
produce UCs but are not currently offering a commercial product line. Those companies
are CCR Corporation (Japan), Panasonic Industrial (Japan), and SAFT (France). Due to
current low levels of production the prices for UCs are still very high. However,
manufacturers claim these prices are largely due to economies of scale and, they are
expected to drop drastically as production levels increase. Future prices are expected to
be very competitive with current HEV battery prices.
Chapter 3
THE BATTERY
While the UC is a relatively new device, the battery has been the electric energy
storage device of choice for centuries. Ever since the first useable battery was invented
around 1799 by Alessandro Volta the world has been using batteries. More recently there
have been several advances in battery technology due to Electric and Hybrid Electric
Vehicles hitting the market. The lead acid battery was invented by Gaston Plante in 1859
and to this day has been the predominant battery used in the automobile industry.
However, advances in battery technology have produced some competition to the lead
acid battery. Both Nickel-Metal Hydride (NiMH) and Lithium-Ion batteries have now
been used in many Electric and Hybrid Electric Vehicles. These batteries have many
benefits over the lead acid battery, but there also are disadvantages.
The Lead-Acid battery is, by far, the cheapest and simplest battery technology
currently on the market. However lead acids also have low energy density, short
longevity, and poor low temperature performance [22]. The other main competitors, the
Nickel-Metal Hydride battery and the Lithium-Ion battery both have a much higher
energy density with the Lithium-Ion battery possessing the highest energy density of the
three. The main drawbacks to these advanced batteries are high cost and added
complexity [22]. Nickel-Metal Hydride has a high self-discharge rate and is challenging to
recharge properly [22]. Lithium-Ion is also complicated to charge due to safety hazards
5
6
inherent in the chemical makeup of the battery [22]. This added complexity adds even
more to the overall cost of the energy storage system. Shown below, in Table 3.1, is a
chart depicting the status of batteries in competition with Lead-Acid for the Electric and
Hybrid Electric Vehicle market.
Table 3.1 – Status of Batteries in Competition for EV and HEV Market Share [22]
System
Energy
Power
Cycle
Cost
Density
Density
Life
(Wh/kg)
(W/kg)
(US$/kWh)
Acidic aqueous solution
Lead/Acid
Alkaline Aqueous
35-50
150-400
500-1000
120-150
Nickel/Cadmium
Nickel/Iron
Nickel/Zinc
Nickel/Metal Hydride
Aluminum/Air
Iron/Air
Zinc/Air
Flow
40-60
50-60
55-75
70-95
200-300
80-120
100-220
80-150
80-150
170-260
200-300
160
90
30-80
800
1500-2000
300
750-1200+
?
500+
600+
250-350
200-400
100-300
200-350
?
50
90-120
Zinc/Bromine
Vanadium Redox
Molten Salt
70-85
20-30
90-110
110
500-2000
?
200-250
400-450
Sodium/Sulfur
Sodium/Nickel Chloride
Lithium/Iron Sulfide
Organic/Lithium
150-240
90-120
100-130
230
130-160
150-250
800+
1200+
1000+
250-450
230-345
110
80-130
200-300
1000+
200
Solution
Lithium-Ion
The United States Advanced Battery Consortium (USABC) has overseen research
in battery systems since its creation in January of 1991[23]. This organization has
published a series of objectives, as shown below in Table 3.2 and these objectives have
driven much of the research in the advanced battery community.
Table 3.2 – USABC Goals for Advanced Batteries for EVs [23]
7
Parameter (Units) of fully
burdened system
Power Density (W/L)
Specific Power – Discharge 80%
DOD/30 sec(W/kg)
Specific Power – Regen, 20%
DOD/10 sec(W/kg)
Energy Density – C/3 Discharge
Rate(Wh/L)
Specific Energy – C/3 Discharge
Rate(Wh/kg)
Specific Power / Specific Energy Ratio
Total Pack Size (kWh)
Life (years)
Cycle Life – 80% DOD (Cycles)
Power & Capacity Degradation
(% of rated spec)
Selling Price – 25,000 units @
40 kWh ($/kWh)
Operating Environment (ºC)
Normal Recharge Time
High Rate Charge
Continuous discharge in
1 hour – No Failure
(% of rated energy capacity)
Minimum Goals for Long
Term Commercialization
460
300
Long Term
Goal
600
400
150
200
230
300
150
200
2:1
40
10
1000
20
2:1
40
10
1000
20
<150
100
-40 to +50
20% Performance Loss
(10% desired)
6 hours (4 hours Desired)
20-70% SOC in <30min @
150W/kg (<20min @
270W/kg Desired)
75
-40 to +85
3 to 6 hours
40-80%
SOC in 15
min
75
3.1: Failure in Lead Acid Batteries
As mentioned above, one of the drawbacks to the lead acid battery is their short
lifetime, and many different failure modes have been identified [24,25,26,27]. However, lead
acid batteries used in HEVs primarily fail due to a process called sulfation [24,26,27,28,29,30].
A lead acid battery consists of three major components, the positive plate, the negative
plate, and the electrolyte. The positive plate is constructed mainly from lead dioxide
(PbO2). The negative plate is constructed mainly of lead (Pb). And the electrolyte
8
consists mainly of sulfuric acid. In a new and fully charged battery undergoing a
discharge, lead (Pb) and lead dioxide (PbO2) react with the sulfuric acid (H2SO4) to
produce electricity, lead sulfate (PbSO4), and water (H2O) according to the following
chemical equation.
Pb + PbO2 + 2H2SO4 ↔ electricity + 2PbSO4 + 2H2O
When the battery is charged back up the lead sulfate and water are broken up into sulfuric
acid, lead, and lead dioxide. Unfortunately, during every cycle a very small amount of
lead sulfate is not able to be broken up, and it hardens on the negative plate decreasing
the available surface area of the plate which decreases the battery capacity. If the battery
is discharged or charged at a high rate or left in a state of discharge for any length of
time, sulfation is greatly increased [24,26,27,28,29,30]. During a high rate discharge the rate of
lead sulfate creation exceeds the rate at which it can penetrate the surface of the plate and
the crystals begin to build on top of one another. This makes it difficult to completely
reform the built-up lead sulfate back into lead (due to a reduction in surface area of the
plate), and it leaves a certain amount of lead sulfate on the plate and sulfation results. If,
a high rate charge is delivered to the battery after a high rate discharge, a process known
as hydrogen evolution can occur. Finally, the largest contributor to sulfation in lead acid
batteries is the usual partial state of charge required in HEVs, i.e., the state of charge is
usually less than 100%. This translates into lead sulfate being left on the plates for an
extended period of time which also results in sulfation of the negative plate.
If a lead acid battery is kept at 100% SOC, and high rates of charge or discharge
are avoided, then its lifetime begins to compete with those of more advanced batteries. A
9
large amount of research is being conducted to prevent sulfation, but one of the most
promising is to add an UC to create a Hybrid Energy Storage System (HESS).
Chapter 4
PRESENT UC – BATTERY HYBRID ENERGY STORAGE
SYSTEMS
The term HESS simply means an energy storage system containing more than one
energy storage devices, in this case, a battery and an UC. The concept is generally
successful because it exploits the strengths and compensates for the weaknesses of each
device. When these systems are used, the batteries can be designed for higher energy
density at the expense of lower power density. Conversely, the UC can be designed for
higher power density at the expense of lower energy density. Previous studies [2,3,4,5,6,7,8],
most of which were based on a simple parallel combination as in figures 1 and 4, have
shown that an UC – Battery combination does indeed increase the performance.
Because of its lower resistance, the UC is able to shield the battery from at least a
portion of the current pulses and thus extend the battery lifetime somewhat [14]. However,
the battery is still exposed to most of the pulse, so the improvement is not as great as
desired. To extend the lifetime, it is necessary to reduce the ionization within the battery,
a phenomenon that is not present in the UC.
10
11
+
IO
i uc
iB
VRUC
+
VUC
-
+
RC
+
R ESR
+
UC
VB
R OV
VRB
VO
-
B
-
-
Figure 4.1 - Equivalent circuit of a parallel battery/UC.
If the parameters in figure 4.1 can be determined, the circuit equations can be
used to predict the behavior for various types of loads. Figure 4.2 shows the results of a
simulation where the constant current pulse in (a) is applied to the circuit in figure 4.1.
From t0 to t1 the UC current in figure 4.2(b) supplies most of the energy while the
battery current in figure 4.2(c) ramps up slowly. Between t1 and t2 the Battery is actually
recharging the UC. This is due to the difference between the Equivalent Series
Resistance of the UC, RESR, and the combination of the Over Voltage Resistance (Rov) and
the Coulomb Resistance (Rc) in the battery [31]. Rov in figure 4.1 is not an actual resistance
in the usual sense but a term that represents the energy losses required to charge or
discharge the battery. This simulation used a constant value for Rov, but this term actually
is very non-linear[31] with respect to the current and the state of charge. In figure 4.1 we
see that between t0 and t1, VRUC + VUC = VRB + VB, but VRUC ≠ VRB, and VB ≠ VUC.
Therefore, at time t1, where the applied current load drops to 0, current will flow from
UC to B since VUC > VB.
If we assume that the battery voltage (VB), the output voltage (VO), the resistances,
and the capacitances are all constant then a closed form solution can be obtained.
12
A mathematical analysis gives the following for 0 ≤ t ≤ t1,
iUC (t ) =
and
(V
UC 0
)
− VB + I o RB
RB + RESR
e −t /τ
iB (t ) = I o − iUC (t )
(1)
(2)
where :
VUC0 = the initial UC voltage
R B = RC + ROV
τ = C (R B + R ESR )
C = Capacitance of the UC
VO = Output Voltage, constant
V B = Battery Voltage, constant
I o = Output current, constant
iUC (t ) = UC current
i B (t ) = Battery current
Figure 4.2 - Hybrid UC - Battery Simulated Discharge Current Waveforms
13
From t0 to t1, assuming that the UC has been fully charged to the voltage of the
battery, iuc becomes,
iUC (t ) =
I o RB
e −t /τ .
RB + RESR
(3)
From t1 to t2, iuc becomes
iUC (t ) =
where :
(
VUC1 = I o RB e
−T p / τ
(V
UC1
− VB
RB + RESR
)e
−t /τ
.
(4)
)
− 1 + VUC0 (the UC voltage at time t 1 ) (see figure 4.2(d))
T p = t1 − t 0 = Current pulse length in seconds
In figure 4.2(d) notice the UC voltage before the pulse, VUC0, and the UC voltage
immediately following the pulse, VUC1. Immediately after the pulse the UC voltage
begins to rise because the battery is recharging the UC. It continues to rise until VUC =
VB. Naturally, the UC current also reverses direction at t1 when the UC begins to charge.
Chapter 5
THE PROPOSED UC – BATTERY ENERGY STORAGE SYSTEM
A conventional UC-Battery HESS offers some advantages, including an increased
battery lifespan, but there is room for a large improvement. The existing system provides
the battery with a limited shield from the large power draws during a typical driving
cycle, but the battery still sees most of each current pulse. The current system also does
not address the partial state of charge operation that results in sulfation. However, if it
was possible to store more energy in the UC and shield the battery from more of the
power draw and keep the battery at 100% state of charge, then the battery would see an
even greater increase in lifespan. Increasing the size of the UC in the existing system
would decrease the average power drawn from the battery, but the total energy drawn
from the battery would stay the same. Also, the size of the UC will always be limited by
weight and volume restrictions.
These problems can be addressed by using multiple voltage levels within the
HESS, but this is cost prohibitive if expensive DC/DC converters [11] are used. However,
two alternative circuits shown in figure 5.1 avoid the limitations of the conventional
system, and the additional hardware is relatively inexpensive.
14
15
Regen
Boost
+
D
+ UC
VUC
+
VB B
a.
VD
Regen
Boost
S
+
B
V
+ UC B
V
OR UC
+
VD
D
b.
S
Figure 5.1 - Proposed Hybrid Battery – UC ESS
For the circuits in Figure 5.1, the UC can be at a higher voltage level than the
battery and can shield the battery from as much of the pulse as required by increasing the
energy storage in the UC. The battery can now be kept up to 100% state of charge in
order to reduce the sulfation of the negative plate. The switch, S, in these circuits is used
only to bypass D when B needs to be recharged. This means that UC must absorb all of
the regenerative braking pulses since current in this direction is blocked by D. The UC
also supplies all of the boost pulses until VUC decreases to VB. The two circuits differ
only in the position of the diode, which does not change the performance. Circuit 5.1(a)
is beneficial when the battery is used as a power source for auxiliary loads because the
negative terminal of the battery is grounded. Circuit 5.1(b) is beneficial because the
diode’s heatsink can be grounded and control of S is simplified if a semiconductor switch
is used. When the UC is fully charged this configuration can supply power to the load
for a short time without any drain on the battery. If the pulse length, TP, is greater than a
certain time, TPC, the battery supplies current after TPC, and the system then behaves like
the system in figure 4.1, assuming a constant voltage drop across the diode, VD. As an
16
alternative to the switch, other means can be added to charge the battery in which case
the switch can be removed. These charging methods will be discussed in chapter 6.
As noted above, after TPC the system behaves exactly like the system in figure
4.1. The equations become identical at that point except for that VD is subtracted from
the Battery Voltage, VB. Before TPC, the equations are as follows.
dvuc (t )
(4.5a)
dt
t
1
vuc (t ) = vuc0 − ∫ iuc (t )dt (4.5b)
C0
During acceleration, it is assumed that the load is best approximated by a constant power
iuc (t ) = C
pulse, Po, thus,
Po = v o (t ) ⋅ iuc (t ) = Constant
(4.6)
Substituting (4.5b) and vo (t ) = vuc (t ) − iuc (t )R ESR into (4.6)
t


Po = iuc (t ) vuc0 − ∫ iuc (t )dt  − iuc2 (t )R ESR .
0


Dividing through by iUC (t ) and differentiating both sides yields
−
Po diuc (t )
i (t )
di (t )
= − uc − RESR uc
C
dt
i (t ) dt
2
uc
And combining terms gives,
diuc (t )
iuc3 (t )
+
=0
dt
C iuc2 (t )RESR − Po
(
)
(4.7)
Equation (4.7) does not lend itself to a closed form solution. Therefore an iterative
solution is required.
From (4.5a) we can obtain dvuc (t ) =
1
iuc (t )dt .
C
If we let dt = ∆t and k = the iteration number, ∆vuc (t ) = vuc ( k −1) − vuc ( k ) , then
17
vuc ( k ) = vuc ( k −1) −
1
iuc ( k ) ∆t .
C
(4.8)
From (4.6) and since vo = vuc (t ) − iuc (t )RESR ,
Po = vo iuc (t ) = (vuc (t ) − iuc (t )RESR )iuc (t ) .
Multiplying through and reconfiguring the terms gives R ESR iuc2 ( k ) − vuc ( k ) iuc ( k ) + Po = 0
∆t 

Substituting 4.8 yields  RESR + iuc2 ( k ) − vuc ( k −1)iuc ( k ) + Po = 0
C

and solving the quadratic produces
iuc ( k )
∆t 

vuc ( k −1) − vuc2 ( k −1) − 4 R ESR +  Po
C

=
∆t 

2 R ESR + 
C

.
(4.9)
Using equations (4.8) and (4.9) we can run a simulation to predict the actions of the
circuit.
A simulation of the circuit in Figure 5.1(a) is shown in Figure 5.2. A constant
10kW is drawn from the system for 9 seconds. Note that the battery does not supply
power until time TPC and that the UC still supplies most of the power. This pulse
simulates an acceleration boost to the vehicle.
18
Figure 5.2 – Simulated Constant Power Pulse
Notice that the UC provides most of the energy for the acceleration boost in figure
5.2. As compared to using a battery alone, using the proposed UC + B combination does
not translate into fuel savings because the same amount of energy is still being supplied,
but it does translate into an increase in battery lifetime. Because the UC has a rated
lifetime greater than that of a typical vehicle, e.g., 10 years or more[32], the lifetime of the
ESS is essentially limited only by the battery, which is dependant upon its use. An
improvement is now achieved because the number and size of the current pulses
delivered by the battery are reduced.
Another important aspect is the required size of the batteries. A conventional
system needs a higher battery voltage and amp hour capacity than the proposed system.
A typical example might be to use three series connected 12 Volt, 40 Amp Hour
19
automotive batteries. This results in a 36 Volt (nominal) system. The proposed variable
voltage system could utilize a 36 Volt UC along with two 12 Volt 15 Amp Hour
batteries. A typical UC would be similar in size and weight to one 12 Volt, 40 Amp Hour
battery, so the total size and weight is about the same when accounting for the added
electronics.
Another important aspect of this configuration is that the negative current pulse
produced by regenerative braking is completely consumed by the UC due to the blocking
diode. Therefore, the battery no longer needs to be sized to accept the energy developed
during regenerative braking.
Chapter 6
BATTERY CHARGING METHODS FOR THE PROPOSED SYSTEM
To allow the battery in figure 5.1 to be charged, a current path must be provided
to bypass D. There are two basic methods for this. The first, shown in figure 6.1, is to
provide an alternate current path in parallel with the diode, and the second is to provide a
separate charging circuit in parallel with the battery. The first method can be done
simply with either a mechanical or a solid state switch.
UC
or
D
B
Figure 6.1 - First Battery Charging Method
Providing a current path in parallel with the diode is simple, but it introduces
some inconveniences. The main issue is that, during charging, the UC must be at the
same voltage level as the battery. Therefore, before the switch is closed the UC must be
discharged down to the battery voltage. To avoid excessive cost, the bypass switch
should be designed to carry only the battery charge current, which is much lower than the
boost and regenerative braking surge currents. This allows an inexpensive automotive
relay to be used. However, this method does not allow for the shielding of the battery
during charging. It is also not possible to charge the battery during a regenerative
braking pulse. Another aspect of the system that isn’t shown in figure 6.1 is another relay
20
21
used for a rare UC discharging function. When the vehicle is at zero velocity there is no
way to discharge the UC and in turn charge the battery. Therefore, a second relay in
series with a resistor is needed to bleed down the voltage of the UC. This is installed in
parallel with the first relay.
A second method shown in Figure 6.2 [1], indicates a Buck Converter configured
in two different ways. Basically, these two circuits differ in the same way as in Figure
5.1, i.e., the position of the diode, D1, has been changed. This also effects the
configuration of the charger. In circuit (a) the charger works by pulse width modulating
(PWM) Q1 at a high frequency such as 100kHz. When Q1 is on, current flows through
Q1, L, and B. When Q1 turns off, the energy stored in L is dumped into B through D2.
I
I
Charger
D1
UC
B
OR
L
D2
B
(a)
D2
L
Q1
UC
Q1
D1
Charger
(b)
Figure 6.2 Second Battery Charging Method
When Q1 is on in circuit (b), current flows through B, L, and Q1. When Q1 turns off, the
energy stored in L is dumped into B through D2. The advantage of both of these methods
is that the UC does not have to be discharged in order to charge the battery. Therefore,
the battery is still shielded by the UC during charging and the battery can be recharged
during a regenerative braking pulse. However, this method requires slightly more space
and cost than the first method.
22
Both charging methods are equally viable, but a determination as to which is
better depends on many factors. Comparatively, both methods have advantages and
disadvantages.
Method 1:
•
+ The hardware is simple, cheap, and relatively small
•
- The necessary control strategy more complicated
•
- The UC must be discharged in order to charge the battery
•
- The battery is unshielded during a charge
•
- The battery cannot be charged during a regenerative braking pulse
Method 2:
•
- The hardware is somewhat larger and more expensive
•
+ The necessary control strategy is relatively simple
•
- The UC voltage must be charged above the battery voltage in order to
charge the battery
•
+ The battery is shielded during charging
•
+ The battery can be charged during a regenerative braking pulse
Upon first glance it looks as if Method 2 is the better method, but this does not take the
HEV control strategy into consideration, which is studied in the next chapter. A few
observations are helpful to this analysis.
•
The only time that battery charging is efficient is when the ICE is operating in an
efficient area. The UC needs to be discharged down to the level of the battery
before charging can commence in method 1, but the energy in the UC is used to
23
accelerate the vehicle so that the UC is already discharged when the ICE takes
over. The ICE now enters its efficient area, which is needed for battery charging.
Therefore, this disadvantage becomes minor. However, it still exists due to other
situations which can occur but are probably infrequent.
•
During a charge using method 1 the battery is unshielded because the UC has
previously been discharged. However, as mentioned above, battery charging
occurs only in an efficient area of the internal combustion engine. Normally, the
UC has already been discharged due to acceleration before charging so the battery
is already unshielded. Therefore, this disadvantage also becomes minor.
•
In method 1 the battery cannot be charged during a regenerative braking pulse.
This issue can be minimized by considering the total time that the vehicle is in
this mode. During some driving cycles there is a fairly high ratio of regenerative
braking to total time, but on average regenerative braking does not occupy a very
large portion of the driving cycle.
•
Because the battery is not used as extensively as it would be without the UC, the
battery does not need to be charged at every opportunity. Therefore, with both
charging methods, the charging control strategy can wait to charge the battery
when it is most advantageous. For instance, the control strategy in the first
method can wait until the UC is already discharged unless the batteries are in a
desperate need of charging. Similarly, the control strategy in the second method
can wait until the UC is at a voltage level above that of the battery.
The main item left to consider is the relative cost, size, and complexity. Method 2
is more expensive, but probably not by much. Method 1 is more complicated to control,
24
but the hardware itself is less complicated. Finally, method 1 is slightly smaller in size,
but again not by much. In summary, method 1 has a slight cost advantage, but method 2
simplifies the control strategy. The advantage of the slightly lower cost is clear, but
without more operating experience, the exact advantage of the simplified control has not
yet been determined. This research has focused primarily on method 1, but more
experience with actual driving cycles may later indicate a preference for method 2.
Chapter 7
PERFORMANCE CHARACTERIZATION METHODS
The design of an ESS is never complete without adequate lab testing. To perform
such tests, most of the automobile industry uses a programmable Vehicle Load Simulator
(VLS), such as the Aerovironment ABC-150 Power Processing System[33]. This device
can be programmed to cycle the ESS through a prescribed drive cycle that is
representative of average driving in a described setting, such as city driving or highway
driving. The VLS is programmed by using a script file which is limited to simple
commands, normally a series of various loads and/or energy pulses. The ABC-150 itself
is not able to implement complex dynamic vehicle models. However, it does provide a
serial data link called the BytePipeTM[33], which enables the VLS to interface with another
device. The addition of a microcontroller to the VLS, via this BytePipeTM, enables it to
incorporate a dynamic model of a vehicle which provides certain advantages.
The test system, shown below in figure 7.1, is centered on the Primary
Microcontroller (PM), which functions as the main control system. It controls the
functions of the charging and discharging relay which in turn is determined by the battery
charging algorithm. It also simulates the vehicle using a series of dynamic equations and
“driver” potentiometer inputs that simulate the accelerator and brake. The results from
this simulation provide the different variables that are used to calculate the required
current from the VLS.
25
26
IRegenerate
10 Ohm
100W
UC
Ultra
Capacitor
IAccelerate
+
VD1
Charging
Relay
D1
-
145 Farads,
42V max
ABC-150
Vehicle
Load
Simulator
B1
Discharging
Relay
12Vdc
RS-232
B2
12Vdc
IUC
IB
VB2
Secondary
Microcontroller
Simulated Velocity
and ESS Voltages
VB1
Primary
Microcontroller
Record
Switch
VUC
RS-232 ABC-150
Current Command
ABC-150
Remote
Operator's
Station
Playback
Switch
Accelerator
Brake
PC
Monitor
ESS Voltages, IUC ,IB , IT ,and Velocity
Figure 7.1 - The Proposed Performance Characterization Method
One of these variables is vehicle velocity. The PM transmits the simulated
velocity as well as the ESS UC and battery voltage levels to the Secondary
Microcontroller (SM), via a CAN bus, and these are displayed on the PC monitor. The
PM also transmits the required current for the ESS to the ABC-150 Remote Operator’s
Station (ROS) via the BytePipeTM.
The main responsibility of the SM is to gather information to be displayed on the
PC Monitor. As seen in figure 7.1 the SM reads the currents of the UC and the battery
via two current sensors. The SM also receives CAN bus transmissions from the PM.
These transmissions consist of the vehicle velocity and ESS voltage levels. The vehicle
27
velocity and the ESS current and voltage levels are then transmitted to the PC Monitor
via a serial connection where they are displayed graphically.
As discussed above, the PM transmits the required ESS current to the ROS. The
ROS runs a script file that constantly monitors the BytePipeTM interface and responds by
commanding the required current from the ABC-150 every time a new command is sent.
The two variable resistors in figure 7.1 model the Accelerator and the Brake.
These potentiometers provide voltage signals that function as the “driver” input to the
simulator. With these inputs the test engineer is able to “drive” the vehicle and see the
real-time vehicle speed as well as the real-time current being taken from or delivered to
the different portions of the ESS. Also shown in figure 7.1 are two switches. The record
and playback switches are inputs to the PM and allow the user to record a driving cycle
and play it back. With these two switches there are three operable modes; Normal,
Record, and Playback.
7.1 The Vehicle Simulation
As outlined above, the PM simulates the vehicle using its algorithm. This
simulation is based on a dynamic vehicle equation including a series of constants that
describe a specific vehicle. A set of conventional equations that describe the vehicle is
listed below.
TQ =
F ⋅ RW
12 ⋅ GR
(7.1)
2  88 
CD ⋅ RHO ⋅ AR ⋅ (V + WV ) ⋅  
 60 
DA =
64.34
2
(7.2)

V 

F = DA + CRO ⋅ 1 +
W ⋅ cos(GA) + W ⋅ sin (GA)
 100 

(7.3)
28
Where:
AR = Frontal Area in square feet
CD = Drag Coefficient
CRO = Coefficient of Rolling Resistance
GA = Grade Angle in radians
RHO = Air Density in pounds per cubic foot
RW = Rolling Radius in inches
WV = Wind Velocity in miles per hour
W = Vehicle Weight in pounds
GR = Gear Ratio
DA = Aerodynamic Drag in pounds
F = Force in pounds
TQ = Motor Torque in foot pounds
V = Vehicle Velocity in miles per hour
Equation 7.4 shows the total force on the vehicle,
FTotal = FAcceleration + F = W ⋅
dV
+F
dt
(7.4)
Ignoring wind velocity, converting speed to meters per second, weight to
kilograms, and force to newtons as well as substituting equations 7.2 and 7.3 into
equation 7.4 yields,
FTotal = (.4536) ⋅W
dV
+ (.744) ⋅ CD ⋅ RHO ⋅ AR ⋅V 2 +
dt
(4.448) ⋅W ⋅ CRO ⋅ (1 + (.02237) ⋅V ) cos(GA) + (4.448) ⋅W ⋅ sin (GA) (7.5)
and regrouping terms gives,
FTotal = (.4536 ⋅ W )
dV
+ (.744 ⋅ CD ⋅ RHO ⋅ AR )V 2 + (.0995 ⋅ W ⋅ CRO ⋅ cos(GA))V +
dt
(4.448 ⋅ W ⋅ CRO ⋅ cos(GA) + 4.448 ⋅ W ⋅ sin (GA)).
If,
A = .4536 ⋅ W
B = .744 ⋅ CD ⋅ RHO ⋅ AR
C = .0995 ⋅ W ⋅ CRO ⋅ cos(GA)
D = 4.448 ⋅ W ⋅ CRO ⋅ cos(GA) + 4.448 ⋅ W ⋅ sin (GA)
(7.6)
29
Then
FTotal = A
dV
+ BV 2 + CV + D.
dt
(7.7)
FTotal is determined by the inputs from the simulated accelerator and brake. Negative
values are for braking and positive values are for acceleration. Equation 7.7 must now be
solved for the unknown velocity, V. Solving for
dV
yields,
dt
dV FTotal B 2 C
D
=
− V − V− .
dt
A
A
A
A
(7.8)
Solving equation 7.8 requires an iterative method since it cannot be solved in closed
form. We have selected the Modified Euler Method using the second estimation as the
solution[34]. This method is outlined below.
dV
from (7.8) where V = V(1t − ∆t ) is the previous solution for V,or the
dt
initial velocity if this is the first iteration.
Step 1: Find
Step 2: Solve for V(t0+ ∆t ) = V(1t − ∆t ) +
dV
⋅ ∆t , where V = V(1t − ∆t ) and ∆t is the time
dt
between iterations.
dV
from (7.8), where V = V(t0+ ∆t ) .
dt
 dV(1t − ∆t ) dV(t0+ ∆t )

+
1
1
dt
Step 4: Solve for V(t + ∆t ) = V(t − ∆t ) +  dt
2


Step 3: Find


 ⋅ ∆t .



V(1t + ∆t ) is the solution for the velocity at the next iteration or time t + ∆t . After step 4,
loop back to step 1 and start the next iteration.
30
To find the value of FTotal, we need to select a maximum vehicle speed, VMax, and
solve for FTotal at VMax when the vehicle is not accelerating. So repeating (7.7) with V =
VMax and
dV
= 0 gives,
dt
2
FTotal Max = BV Max
+ CV Max + D .
(7.9)
This solution results in the maximum value of force that can be used in the above
equations. The brake input should be scaled so that, − FTotalMax ≤ FBrake ≤ 0 . The
accelerator input should be scaled so that, 0 ≤ FAccelerator ≤ FTotalMax . The total force is then
calculated by equation (7.10).
FTotal = FBrake + FAccelerator
(7.10)
7.2 HEV Control Strategy
This section describes the control strategy used to determine the split between the
internal combustion engine and the electric motor. There are many different control
strategies used for HEVs, but the one that has been implemented in this system is similar
to that in [35]. This rule-based strategy uses five operating modes, Electric Motor Only,
Internal Combustion Engine (ICE) Only, Power-Assist (a combination of the motor and
engine), Recharging, and Regenerative Braking. The Recharging Mode and the ICE
Only Mode can occur simultaneously.
There are three rules used to determine the proper operating mode for the vehicle:
Rule #1 - Power Split Rule:
The Power Split Rule determines which prime mover is powering the vehicle.
The first part of this rule places the vehicle into either Motor Only Mode or ICE Only
31
Mode. Below the speed / torque ratio at which the ICE can operate efficiently the motor
operates as the only power source to the vehicle. Once the vehicle crosses the engine-on
line (Pe-on) (see the ICE efficiency map in figure 7.2) the ICE takes over and the motor no
longer supplies power to the vehicle. The second part of this rule puts the vehicle into
ICE Only Mode or Power-Assist Mode. Above the speed / torque ratio at which the ICE
can no longer operate efficiently the motor provides an assist in order to keep the ICE
below the motor-assist line (Pm-a)(again see the ICE efficiency map in figure 7.2).
Figure 7.2 – ICE Efficiency Map and The Power Split Rule [35]
Rule #2 - Recharging Rule:
Normally, an HEV operates so that the SOC of the battery is in the range of 55%80% but it is preferable to keep the battery close to 100% SOC in order to minimize
sulfation, which is the most prominent failure mode for Lead-Acid batteries. Therefore,
32
there are two sub-rules within the Recharging Rule, the Non-Compromising Recharge
Rule and the Compromising Recharge Rule.
Non-Compromising Recharge Rule:
If the vehicle is operating in the ICE Only Mode and the battery voltage is below
the Non-Compromising Recharge Limit (currently set to 12 volts for each of the
two batteries), the vehicle will enter the Recharge Mode if the UC does not need
to be discharged, i.e., if its voltage has been discharged to the same level as the
battery. The ICE will still be the only power source moving the vehicle, and it
also will be generating a small amount of extra torque to run the motor as a
generator in order to supply a charge current to the batteries (currently set to 25
amps). When either of the batteries reaches the upper battery limit (currently set
to 13.5 volts) the charge current drops to a trickle-charge current (currently set to
2 amps) until both batteries reach the upper battery limit.
Compromising Recharge Rule:
If the vehicle is operating in the ICE Only Mode and the battery voltage is below
the Compromising Recharge Limit (currently set to 11 volts) the vehicle will enter
the Recharge Mode even if the UC needs to be discharged. As previously
described, when either of the batteries reaches the upper battery limit (13.5 volts),
the charge current drops to a trickle-charge current (2 amps) until both batteries
reach the upper battery limit. If the vehicle is in Recharge Mode and Regenerative
Braking is required, the Recharge Mode is exited and charging is stopped.
Likewise, if the vehicle is in Recharge Mode and Power-Assist Mode or Motor
Only Mode is required, Recharge Mode is exited.
33
If the vehicle enters the Power-Assist Mode or the Motor Only Mode and remains
there long enough so that either battery reaches the Battery Minimum Voltage (currently
set to 10 volts), the vehicle will switch over to ICE Only Mode and recharging will occur.
If the vehicle is still being operated below the Pe-on line or above the Pm-a line when the
batteries have been charged then the vehicle will revert back to the Motor Only Mode or
Power-Assist Mode respectively without performing the trickle charge.
Rule #3: Braking Rule:
During deceleration of the vehicle there are three different situations that can take
place. First, the vehicle can be coasting which means that neither regeneration nor the
friction brake is applied. The second situation is when only regen is applied and the
energy storage system is being charged. In the third situation, the energy storage system
is fully charged and cannot absorb any energy from the generator. In this case only the
friction brakes will be applied.
Motor Only, ICE Only, and Power Assist modes are indicated in figure 7.3. The
placement of the mode boundaries in figure 7.3 depends on the efficiencies and other
specifics of the ICE and the motor, all of which is vehicle dependant. Each boundary is
selected in order to maximize the overall efficiency of the vehicle, and each of the
numbered lines on the map represents a locus of constant ICE efficiency points. The
lower values represent higher efficiency.
7.3 System Software
There are four computing modules within the system that require software. These
are the PM, the SM, the ABC-150 Remote Operator Station (ROS), and the PC Monitor.
34
The details of the software within each of these modules are given below and the actual
code is given in Appendix II.
Primary Microcontroller (PM) Software:
The primary microcontroller requires two separate algorithms that run
simultaneously. The first algorithm, shown in figure 7.3, is responsible for battery
charging and routine message transmission to the ABC-150 ROS.
Start
Is battery
voltage <
minimum?
Set ABC-150 to
Zero Current
Yes
Discharge
UC
Is electrical
force = 0 &
velocity > 0?
Set ABC-150 to
Charge Current
Set ABC-150 to
Zero Current
Turn off
Relay
No
Turn on
Relay
Is battery
voltage < full
charge and
is Regen = 0
and
velocity > 0?
No
Is battery
voltage ≤
BCLNC?
Yes
Yes
Yes
No
No
Is battery
voltage ≤
BCLC?
Yes
Send Required Current
Message to ABC-150
No
Discharge
UC
Yes
Yes
Set ABC-150 to
Zero Current
Turn off
Relay
Is battery
voltage < full
charge and
is Regen = 0
and
velocity > 0?
Does UC
need
discharged?
Set ABC-150 to
Charge Current
No
Turn on
Relay
No
Figure 7.3 – Battery Charging and Routine Message Transmission Algorithm
35
There are two modes of battery charging: a Desperate Battery Charging Mode and a
Normal Battery Charging Mode. The Desperate Battery Charging Mode is used when the
voltage level of the battery has dropped below the minimum threshold limit. Within the
Normal Battery Charging Mode are two sub-modes to determine how the charging will
take place. If the battery voltage has dropped below the Battery Charge Level –
Compromise (BCLC) then the system will compromise the voltage level of the UC, by
reducing it so the battery can be charged. If the battery has not dropped below the BCLC
but has dropped below the Battery Charge Level – No Compromise (BCLNC), the
battery will be charged only if the UC is already discharged to the same voltage as the
battery. In all of these battery charging modes, the system continues charging until either
the batteries have been recharged, a regen occurs, or the vehicle stops.
The second algorithm that is executed on the PM is responsible for the vehicle
simulation and the analog interface. This is shown in figure 7.4.
1
2
Obtain analog values
of the ESS voltages
and the User Inputs
Set Flags for Record,
Playback, or Normal
Start
6
1
3
Select
Powertrain Mode
Transmit Velocity and ESS
Voltages via the CAN bus
4
5
Calculate Velocity
Calculate Forces
Figure 7.4 – Analog Interface and Vehicle Simulation Algorithm
36
This algorithm contains six subsections, each of which are described below.
Subsection 1: Set Flags for Record, Playback, or Normal.
As discussed above, this system has the capability of running in three modes:
Normal, Record, and Playback. Subsection 1 determines the operating mode based on
the user inputs. In Record mode and Playback mode it has the responsibility of storing
and recalling data.
Subsection 2: Obtain Analog Values of the ESS Voltages and the User Inputs.
This subsection is responsible for handling the analog inputs to the
microcontroller. It samples and averages the analog channels for the three ESS voltage
level readings and the accelerator and brake potentiometer inputs. This subsection also
must correct for voltage offsets of the ESS voltage level readings. Before being fed into
the microcontroller, the ESS voltages are reduced to a voltage level suitable for the
microcontroller (5 Volts maximum) using a simple resistor divider circuit. Therefore
these voltage levels must be corrected to their original values by the microcontroller. The
resistor divider circuit is a linear divider so a simple coefficient multiplication suffices.
Data mapping for the accelerator and brake inputs is more complex and requires a nonlinear equation since the desired result is non-linear, but the potentiometer input is linear.
In order to make the accelerator and brake function properly the following correction
equation is used:
Desired Value = 10.3 10 ⋅ Input Value .
(7.11)
Subsection 3: Select Powertrain Mode
The main responsibility of this subsection is to determine which of the following
modes the vehicle is operating in: Internal Combustion Engine(ICE) Only Mode,
37
Regenerative Braking Mode, Motor Only Mode, or Power Assist Mode. To make this
decision the Gear Ratio must first be chosen based on the current state of the vehicle.
Once the Gear Ratio is known the Engine RPMs can be determined. The required torque
can be calculated based on the user inputs from the accelerator and the brake. At this
point the vehicle velocity, engine RPMs, battery charging mode, and required torque are
compared to the ICE Efficiency Map, and the set of rules explained in section 7.2 is used
to determine which mode the vehicle is running in. However, there are two exceptions to
this: the transitions from ICE Only Mode to Motor Only Mode and from ICE Only Mode
to Power Assist Mode are both given a small amount of hysteresis to prevent an
oscillatory state, i.e. hunting.
Subsection 4: Calculate Forces
This subsection contains the algorithm responsible for solving the equations for
the force and required current outlined in section 7.1. The required force of electrical
origin is calculated differently, depending upon which of the powertrain operating modes
the vehicle is in. This force is then used to calculate the required current which must be
drawn from the ESS by the ABC-150. In playback mode the required current is recalled
from memory, and the force is calculated in reverse from the required current. In record
mode the calculated current is saved to memory.
Subsection 5: Calculate Velocity
This subsection calculates the vehicle velocity based on the velocity equations
outlined in section 7.1. In playback mode the vehicle velocity is recalled from memory
instead of being calculated, and in record mode the velocity is saved to memory after
calculation.
38
Subsection 6: Transmit Velocity and ESS Voltages via the CAN bus
Here the data is packaged into a CAN message and transmitted to the SM. The
three ESS voltage levels and the vehicle velocity each occupy their own CAN Message
Object and Identifier. To increase the accuracy of the received data, it is first multiplied
by 1x106 and separated into four bytes before it is sent. The receiver then reverses this
process to obtain the actual value.
Both of these algorithms run simultaneously on the PM. Each algorithm is timed
and occurs on a set frequency unless the algorithm is in a wait state. The first algorithm
is timed to execute 2 times per second and the second algorithm executes 10 times per
second. The second algorithm contains no wait states so it is never delayed. However,
the first algorithm contains multiple wait states. When the battery is being charged the
first algorithm pauses until one of the conditionals become false.
Secondary Microcontroller Software:
The SM is responsible for the collection of data and the transmission of data to the
PC Monitor for display. Data is collected both from the CAN bus and from the analog
inputs from the ESS current sensors. In order to carry out these duties the SM uses three
simultaneous algorithms shown below in figure 7.5.
Algorithm 1 is responsible for the interface to the current sensors. It samples both
inputs 100 times and averages them before updating the variables in memory. These
updated values are sent, along with the ESS voltages and the velocity to the PC Monitor
by algorithm 2. The ESS voltages and the velocity are gathered from the CAN bus by
algorithm 3. This algorithm is also responsible for filtering out any errors due to noise
that might have occurred in the transmission.
39
Algorithm 2
Algorithm 1
Start
Start
Send current data out to
PC Monitor via the USART:
Sample analog
values from the
Current Sensors
Vehicle Velocity, Ultracapacitor Current,
Battery Current, Ultracapacitor Voltage
Battery 1 Voltage, Battery 2 Voltage
Increment Counter
And add samples to
totals
Is
Counter > 99
Algorithm 3
No
Yes
No
Start
Has a CAN
message
been
received?
Divide totals by 100
and save to memory
Divide received
values and filter
out transmission
errors
Yes
Save results to
memory for
USART
transmission
Figure 7.5– Secondary Microcontroller Algorithms
PC Monitor Software:
The PC Monitor software was developed in a graphical programming
environment called StampPlot [36], which is similar to National Instrument’s LabVIEW.
The serial communications port on the computer is used to bring the data into the
program. StampPlot automatically recognizes up to ten comma delimited numbers and
dumps each number into a variable (AINVAL0 – AINVAL9) for use in the program.
Each transmission is completed with a carriage return to signify the end of a time
interval. StampPlot automatically updates the variables AINVAL0 – AINVAL9 after
each transmission. In the code, each variable is tied to one or more objects on the screen.
For this program the following variable assignments are used.
40
AINVAL0 – Not used
AINVAL1 – Vehicle Velocity
AINVAL2 – UC Current
AINVAL3 – Battery Current
AINVAL4 – UC Voltage
AINVAL5 – Battery 1 Voltage
AINVAL6 – Battery 2 Voltage
AINVAL7 – Not used
AINVAL8 – Not used
AINVAL9 – Not used
Along with channeling the data into the appropriate variable, the code also filters out any
values that are out of range for the particular object. This avoids the common occurrence
of noise that causes data errors in transmission but cannot be removed with an analog
filter. Figure 7.6 shows a screen shot of the display developed with the StampPlot
software.
Figure 7.6 – PC Monitor Software screen shot
41
A brief explanation of each part of figure 7.6 is given below, starting at the left in
the top row.
Vehicle Speed Plot: This plot gives the speed of the simulated vehicle in miles per hour.
The indicated span of the x axis is 120 seconds, but this can be changed by using the drop
down list at the top center of the screen. This selection is used to change all plots on the
screen.
Vehicle Speedometer: This object gives a current value of the vehicle speed in miles
per hour by displaying the needle position and presenting a digital readout.
Screen Capture Buttons: These buttons allow the user to capture all or parts of the
display. The first button allows the user to perform a screen capture, and was used to
produce figure 7.5. The second button allows the user to capture all four plots at the
same time. The output is saved as four separate files in the StampPlot data directory. For
convenience, each of the last four buttons can be used to capture a specific plot.
UC Voltage Gauge: The UC Voltage Gauge is a graphical representation of the voltage
level of the UC as well as a digital readout. The white line is the current voltage level of
the UC and the red bar gives a static indication of the nominal voltage level of the
batteries (24 Volts). This allows the user to determine when the UC has been discharged
down to the level of the batteries.
Battery 1 and Battery 2 Gauges: These gauges show the voltage levels of the batteries.
The white line represents the current voltage level of the battery and the red bar gives a
static indication of the nominal battery voltage (12 Volts). Below each of the gauges is
the digital readout of the voltage level indicated in the gauge.
42
UC Current Plot: The UC Current is displayed in amps in the lower left corner of the
screen. This span of the x axis uses 120 seconds, but it also can be changed by using the
drop down list at the top center of the screen.
Battery Current Plot: The Battery Current is displayed in amps in the lower right
corner of the screen. The span of the x axis again is 120 seconds, but can be changed by
using the drop down list at the top center of the screen.
ABC-150 ROS Software:
The ABC-150 ROS software consists of an algorithm that deciphers the messages
being sent to it from the PM, after which it sets the required current from the ABC-150.
The Algorithm is shown below in figure 7.7.
The algorithm starts by waiting for a message to be received from the PM. Once
the message is received, it is broken into two parts: the sign and the magnitude. Two
special cases, 0 amps and ±1 amps, also are accounted for. The ABC-150 is then set to
the proper current, and the algorithm returns to the wait state. For safety reasons, the
current drops to zero if it has been more than 10 seconds since the last message was
received.
43
Start
Read in 2
bytes
Is
byte 1
No
Yes
No
Has a new
message
been
received?
Current is
Positive
Yes
No
Current is
equal to
±(Byte 2)*2
No
Is
Byte 2 = 220?
Current is
Negative
Is
Byte 2 = 250?
Yes
Current is
Zero
Yes
Current is
±1
Set the ABC-150 to the appropriate current
Figure 7.7 – ABC-150 ROS Algorithm
7.4 System Hardware
Not shown in figure 7.1 are the supporting electronics used to process the signals
between the analog portions of the system and the digital portion of the system.
Various driver circuits and signal conditioning circuits are used to interface between
the two portions.
Figure 7.8 shows the circuit that decreases the input voltage by a factor of 11 in
order to bring the voltage down to a level that is acceptable to the microcontroller.
This circuit is used to bring the UC voltage (45 Volts) down to a level that is less than
44
5 volts which is the maximum allowable voltage input to the onboard analog to digital
converter. Two circuits that are identical to figure 7.8, except for the value of R2, are
used for the other two battery sense lines. The values of R2 for these two circuits are
30k and 60k, which yield reduction factors of reduction of 7.67 and 4.33,
respectively. These are used to reduce the total battery voltage (24 volts) and the
battery center tap voltage (12 volts) down to 5 volts.
From Battery
+12VDC
200k
R1
3
20k
R2
U1A
+
4
LM6134
2
-
1 To Microcontroller
11
Figure 7.8 – Battery Voltage Sense signal conditioning
The IR2118 based relay driver circuit is shown in figure 7.9. Because the IR2118
requires a 12V input voltage, the output from the microcontroller must be amplified. A
non-inverting amplifier configuration was used with a gain of three. The output of the
IR2118 is used to drive the base of the VN10LP MOSFET. The MOSFET operates as a
switch to energize the relay coil. The input to the IR2118 is active low, which is required
because of a short boot-up routine that occurs in the microcontroller. During the boot-up
sequence the outputs are held high, meaning an active high device could produce
unwanted triggering. This is important because an unwanted closing of the relay could
damage its contacts since it could possibly short the 24 volt battery to the 45 volt UC. At
the very least this would destroy the relay and could possibly damage the UC. This
circuit is used for both the discharging relay and the charging relay.
45
+12VDC
1N4007
+12VDC
D3
U2A
From
3
Microcontroller +
4
2
-
2
1
LM6134
3
11
2.2uF
20k
10k
COM
HO
R14
M1
VN10LP
IR2118
1
VS
VCC
+12VDC
7
20
C1
R12
R13
__
IN
To Relay
Coil
VB
8
6
U3
Figure 7.9 – IR2118 Based Relay Driver Circuit
The current transducers used to measure the battery and UC current are the LEM
LF505S. This transducer is a closed loop, hall effect transducer with an output ratio of
5000:1. This translates into 20mA output for 100A input, or Iout = Iin / 5000. Given an
input range of ±400A, the output range will be ±80mA. Shown in figure 7.10 is the
circuitry used to adapt the output current from the sensor to the microcontroller. With a
load resistance (R1) of 56 Ohms, this yields (Iin / 5000)*R1 volts at pin 3 of the LF356.
With a ±400A input range, this yields a voltage range at pin 3 of the LF356 of ±4.48
volts. The LF356 voltage follower leaves the voltage level ideally unchanged, but the
voltage divider presented by R2 and R3 divides and offsets the voltage in order to get the
voltage into the range of 0 – 5 volts for the microcontroller. This produces a voltage at
pin 3 of the LM6134 equal to equation 7.11. Again, using the ±400A input range, this
yields a range of 0 - 4.74 volts at the input to the microcontroller.
1  56 ⋅ I in 
Vout = 
 + 2.5 volts
2  5000 
(8.11)
46
Another important feature of this circuit is the protection provided by the second voltage
follower. The voltage supply to the LM6134 is 5 volts, therefore the output voltage
cannot exceed 5 volts. This protects the microcontroller from input currents that exceed
the measurable current level of 450A. Thus, the microcontroller will read any input
currents exceeding 450A as 450A.
+5VDC
+15VDC
U1
From
LEM
3
56
R1
C4
10uF
+
C5
4
LF356
2
+5VDC
1uF
20k
1
R2
R3
3
C7
1uF
U3A
+
4
20k
-
-15VDC
11
LM6134
C6
1uF
2
-
1
To
Microcontroller
11
Figure 7.10 – LEM Current Sensor signal manipulation circuit
Four small telecommunications relays also are used for programming the
microcontrollers but these are not used during normal operation. These are required
because the use of off-board programming switches can cause noise problems.
The remainder of the supporting electronics is self-explanatory and consists of a
few voltage followers, some pull-up resistors and a DC-DC converter. The complete
system schematics are shown in Appendix III.
Chapter 8
EXPERIMENTAL RESULTS
This system was tested in three ways. First, a proof of concept test was done to
test the concept of the UC – battery – diode combination. Second, testing was done on
the system described in chapter 7 to prove the design functionality of battery charging
method 1 (the parallel switch). Finally, brief testing was completed on battery charging
method 2 (the buck converter) to prove the functionality of this alternative.
To perform the first series of tests, the circuit in figure 5.1 was built and tested
without the parallel charging switch. The UC used in this test was a NESSCAP EMHSP0094C0-045R0 module which is rated at 94 Farad, 45V. The battery pack consisted of
(2) 50Ah (c20) Exide 34DC-48 12V valve regulated lead acid (VRLA) batteries, and the
diode was an International Rectifier 1N4049. The Aerovironment ABC-150 Power
Processing System was used generate the required current pulses, and the test currents
were measured using two LEM LF-505S Current Transducers. Figure 8.1 shows the
results of a series of 300A / 5 sec. boost and regen pulses. For the energy level of this
pulse the UC is large enough to both deliver and store 100% of the energy, and the
battery remains inactive. Note that the regen current is not quite constant for the entire 5
sec., i.e., the amplitude begins to drop before the end of the pulse. This is because the
ABC-150 maximum voltage is limited to 45V to protect the UC, and 45V is reached
before the end of the pulse. Similar protection would be provided on an HEV.
47
48
Figure 8.1 - Results for 300 A / 5 sec. test currents
If the pulses are increased to 8 seconds as in figure 8.2, the battery begins to
deliver energy during the final portion of the boost pulse. As the pulse width increases,
the battery supplies more of the boost energy, but it will never consume any regen energy
due to the blocking diode. Each regen pulse starts to decrease before the end of the pulse,
as discussed previously for figure 8.1.
49
Figure 8.2 - Results for 300A / 8 sec. test currents
For idle-stop operation where the ICE is turned off, the ESS must provide energy
to run the auxiliary loads and then re-start the engine. Because of the constant idle-stop
power specification, these tests where performed using constant power pulses as an
alternative to the constant current pulses in figures 8.1 and 8.2. The test results in figure
8.3 show a 1.5kW, 30 sec. load during idle-stop, followed by a constant start up pulse of
5.2kW, 6sec. After the ICE has reached an efficient operating speed, in an actual HEV,
the generator would be used to recharge the UC. Therefore this test was completed by
using regen energy to bring the UC back up to 100% SOC. This is represented by the
2kW, 25 sec. recharge pulse and the 5kW, 6 sec. regen pulse, respectively.
50
Figure 8.3 - Results for the Idle-Stop Test
Next, the system in figure 7.1 was constructed using two Phytec development
boards with Siemens C505CA Microcontrollers[37] and two TYCO, VF7-41H11
automotive relays, along with the components used for the proof of concept tests. The
supporting electronics were discussed in chapter 7, and complete schematics are included
in Appendix III. The simulated vehicle was test driven a number of times in order to
demonstrate all of the HEV control strategy modes; Electric Motor Only, Internal
Combustion Engine (ICE) Only, Power-Assist, Recharging, and Regenerative Braking.
In order to demonstrate these capabilities in a timely manner, very weak batteries were
used so that less time would be needed to discharge and charge. This allows the results
51
of one test period to be displayed on one graph as shown in figures 8.4 – 8.8. Figure 8.4
shows a screen shot of the software that was developed using StampPlot Pro[36] from
Selmaware.
Figure 8.4 – StampPlot Pro screen capture.
For discussion purposes the graphs were captured separately for greater detail and
are shown below in figures 8.5. The simulated drive starts in Motor Only Mode at zero
speed, and by the time the vehicle reaches the speed at which it switches to ICE Only
Mode, the UC has discharged to the level of the battery. Thus, charging of the battery
can occur without further discharging the UC.
52
Figure 8.5 – Simulator results to demonstrate battery charging and ICE Only Mode.
From figure 8.5 we can see that the vehicle begins to charge the batteries at about
10 seconds into the test. Notice that since the UC and battery are now shorted together,
the UC is also being charged. However, the energy into the UC is relatively small
because the net voltage increase is very small. During the first 15 seconds of the battery
charge time the system is charging at 20 amps, and then at 25 seconds the system moves
to a trickle charge current of 2 amps. However, due to the difference in the ESR, of the
UC and battery, the ESR voltage drops are not the same for 2A and 20A. Therefore, at
25 seconds the charge current slowly drops off until it reaches the system charge current
of 2 amps at about 72 seconds. It then continues to trickle charge the batteries until they
have reached 100% SOC.
53
Figure 8.6 shows results that demonstrate the Power Assist Mode of operation.
As before, the test starts with the vehicle in Motor Only Mode as it accelerates from zero
speed. At about 9 seconds into the test the vehicle enters ICE Only Mode. Then at about
29 seconds the system enters the Power Assist Mode. Note that the current of the UC
goes negative, and shortly thereafter the battery current goes negative. Then at 56
seconds the battery current goes positive. This is because one of the batteries has reached
its lower voltage limit set by the system. At 56 seconds the relay has closed and the
charge current has been applied; however, the battery does not see a positive current until
62 seconds due to the voltage mismatch between the UC and battery. After a short
charge time, due to the choice of weak batteries, the batteries are charged again. Note
that in Power Assist and Motor Only Modes the battery charging does not enter the
trickle charge portion of the charging algorithm. At the completion of the charge
Figure 8.6 – Simulator results to demonstrate Power Assist Mode.
54
time, 84 seconds into the test, the UC current again goes negative, and the battery current
follows shortly thereafter. This cycle will continue until Power Assist Mode is exited by
deceleration. While the UC and battery are being used to supply power to the vehicle the
system is in Power Assist Mode, but when the battery is being charged by the vehicle, the
system is in ICE Only Mode.
The Motor Only Mode is shown in figure 8.7. At 24 seconds into the test the
system switches into ICE only mode and charges the battery. Then at about 50 seconds
the system switches back into motor only mode. This cycle will continue until the
vehicle accelerates enough to break into ICE only mode or until the vehicle decelerates.
Note again, that this cycle occurs at high frequency for demonstration purposes only due
to the choice of weak batteries.
Figure 8.7 – Simulator results to demonstrate Motor Only Mode.
55
Figure 8.8 shows an array of different mode executions. This simulates a more
complex driving cycle than those used to show the different operating modes. However,
the important thing here is the demonstration of the regenerative braking. At 77 seconds
and 98 seconds notice that the UC current rises to about +300 amps as the vehicle is
decelerating. This is the regenerative braking mode of operation. Also notice that at 103
seconds the slope of the velocity changes, and the UC current returns back to zero. This
represents coasting of the vehicle.
Figure 8.8 – Simulator results to demonstrate regenerative braking.
For the third test, a version of the parallel battery charging system (method 2) was
constructed as shown in figure 6.2a. The system was implemented with a buck regulator
operated in the current control mode using the TL494MJ PWM Controller as shown in
figure 8.9.
56
+15VDC
+
U3
LF356
15k
100k
+15VDC
L1
40uH
B
D2
To ABC-150
+15VDC
D3
8
UC
-15VDC
15k
100k
-15VDC
1k
IR2118
3
7
D4
50 Ohm
D1
18V
6
2M
10k
2
+15VDC
1k
+15VDC
.001uF
2.2uF
20 Ohm
.1uf
20k
1
.47uF
Q1
.33uF
5
16
12
8
.1uF
11
6
9
TL494MJ
10
1
2
+5VDC
1.5uF
10k
3
15
2M
2M
.33uF
.33uF
14
13 7
20k
10k
20k
4
1.5uF
Figure 8.9 – Buck Regulator Parallel Battery Charging System
The control in figure 8.9 is handled by the TL494MJ controller. This controller
drives the IR2118 IGBT Driver. The IR2118 functions as a current amplifier since the
TL494MJ is only capable of sourcing low currents and higher current pulses are required
because of the input capacitance of the IGBT. The IR2118 drives the IGBT on and off as
directed by the TL494MJ. The switching of the IGBT and general behavior of the power
components are discussed in chapter 6. This converter uses two feedback loops, one for
voltage control and one for current control. Since B is not connected to the ground in this
version, the LF356 operational amplifier is used to scale the battery voltage and shift it to
the ground reference. The LF356 is then connected to the voltage sense input (pin 1) of
57
the TL494MJ. The current loop is implemented by sensing the current in the inductor
and routing the signal into the current sense input (pin 16) of the TL494MJ. The passive
components connected to the TL494MJ serve to set the current and voltage references,
the switching frequency, and the frequency compensation for the two feedback loops.
Once constructed, the results were successful and sample waveforms are shown in figure
8.10. As stated earlier, the benefit of this method is that the battery can always be
charged as long as the UC voltage is at least about 3V above the battery voltage. This
means charging can continue during boost or regen.
Figure 8.10 – Results of the Parallel Battery Charger
Trace 3 in Figure 8.10 shows the inductor (L1) current of the regulator which was
set at 10 amps. The L1 current, iL1, is continuous, but the actual charge current through B
58
is a PWM pulse train. When Q1 is on, iL1 flows through B and Q1. When Q1 is off, iL1
flows through D1 and D2. It can be seen that this current rises to 10 amps after the boost
pulse has removed a certain amount of energy from the battery. As the UC is recharged,
the charge current slowly approaches 10 amps due to the regulator only being able to
supply full current when the UC voltage is sufficiently higher than the battery voltage.
The charge current goes negative while the battery is providing boost to the vehicle
because the reverse parallel diode, D4, of the IGBT and the inductor provide a path in
parallel with the blocking diode, D1. This allows current to pass in the reverse direction
through L1.
Chapter 9
CONCLUSION
The further development of automobiles is important due to rising fuel costs.
While many advances in the automotive industry have taken place since the invention of
the internal combustion engine, it is clear that the advancement presented by the HEV is
one of the most important. This research has focused on an advancement in the energy
storage medium of the HEV, and it has focused on three areas:
1. The development of an improved Hybrid UC – Battery Energy Storage System
where the UC can be charged to a higher voltage than the battery, thus shielding
the battery from most of the current pulses. This increases the life of the battery.
2. The development of a battery charging scheme for the system in 1.
3. The development of an improved method for modeling the performance of the
energy storage systems for HEVs.
A simple scheme has been developed and tested for a variable voltage level energy
storage system, and improvements have been observed. Previous UC – Battery Energy
Storage Systems have been shown to be inadequate for extending the life of the battery
due to their limited current shielding and their inability to keep the battery close to 100%
state of charge. This improved system should drastically reduce the amount of sulfation
in lead acid batteries, which is the primary failure mode in hybrid electric vehicle
59
60
applications. Two battery charging schemes for the improved system have been
explored, and the advantages and disadvantages of each have been discussed. An
improved method for performance characterization has been developed and was used to
test the improved system in the lab. This system uses a mathematical model of the
vehicle to represent the dynamics and this can be tailored to a wide range of vehicles.
Several tests have shown the proposed system to be advantageous, and the
feasibility has been determined using the newly developed performance characterization
method. The advancement has been shown to be a real possibility for the future of hybrid
electric vehicles, but widespread application will depend upon further reductions in the
cost of UCs.
References
[1] T. Stuart, “Circuit with a Switch for Charging a Battery in a Battery Capacitor
Circuit,” Patent Application, The University of Toledo, 2004.
[2] J. Miller, P. McClear, and M. Cohen, “Ultracapacitors as Energy Buffers in a Multiple
Zone Electrical Distribution System,” Maxwell Technologies Inc. White Paper,
www.maxwell.com.
[3] D. Napoli, et al., “Design and Testing of a Fuel-Cell Powered Propulsion System
Supported by a Hybrid UC-Battery Storage,” SAE Journal, pp. 151-160, 2004.
[4] R. Dougal, S. Liu, and R. White, “Power and Life Extension of Battery-Ultracapacitor
Hybrids,” IEEE Transactions on Components and Packaging Technologies, Vol. 25, No.
1, March 2002.
[5] F. Gagleardi, et al., “Experimental Results of on-board Battery-Ultracapacitor System
for Electric Vehicle Applications,” IEEE, 2002.
[6] J. Miller and S. Butler, “Diesel Engine Starting Using Battery-Capacitor
Combinations,” SAE Journal, 2001.
[7] C. Holland et al., “Experimental characterization of hybrid power systems under pulse
current loads,” Journal of Power Sources, Vol. 109, pp. 32-37
[8] X. Yan, D. Patterson, “Novel power management for high performance and cost
reduction in an electric vehicle,” Renewable Energy, Vol. 22, pp. 177-183, 2000.
[9] R. Schupbach, et al., “Design Methodology of a Combined Battery-Ultracapacitor
Energy Storage Unit for Vehicle Power Management,” IEEE 2003 Power Electronics
Specialists Conference (PESC), June 16th, 2003, Acapulco, Mexico.
[10] A. Trippe, A. Burke, and E. Blank, “Improved Electric Vehicle Performance with
Pulsed Power Capacitors,” Proceedings of the ESD Environmental Vehicles Conference
and Exposition, Dearborn, Michigan, 1995.
[11] J. Dixon and M. Ortuzar, “Ultracapacitors + DC-DC Converters in Regenerative
Braking System,” IEEE AESS Systems Magazine, August 2002.
61
62
[12] A. Burke, “Ultracapacitors: why, how, and where is the technology,” Journal of
Power Sources, Vol. 91, pp. 37-50, 2000.
[13] J. Mierlo, P. Bossche, and G. Maggetto, “Models of energy sources for EV and
HEV: fuel cells, batteries, ultracapacitors, flywheels, and engine generators,” Journal of
Power Sources, Vol. 128, pp. 76-89, 2003.
[14] R. Dougal, L. Gao, S. Liu, “Ultracapacitor model with automatic order selection and
capacity scaling for dynamic system simulation,” Journal of Power Sources, Vol. 126, pp.
250-257, 2003.
[15] A. Chu, P. Braatz, “Comparison of commercial supercapacitors and high-power
lithium-ion batteries for power-assist applications in hybrid electric vehicles I. Initial
Characterization,” Journal of Power Sources, Vol. 112, pp. 236-246, 2002.
[16] EPCOS, www.epcos.com
[17] ESMA, www.esma-cap.com
[18] Tavrima, www.tavrima.com
[19] NESSCAP, www.nesscap.com
[20] Maxwell Technologies, www.maxwell.com
[21] PowerSystems, www.powersystems.co.jp
[22] D. Rand, R. Woods, and R. Dell, Batteries for Electric Vehicles, Research Studies
Press Ltd., Taunton, England, 1998.
[23] See, USABC at www.uscar.org.
[24] L.T. Lam, N.P. Haigh, C.G. Phyland, A.J. Urban, “Failure mode of valve-regulated
lead-acid batteries under high-rate partial-state-of-charge operation,” Journal of Power
Sources, Vol. 133, pp. 126-134, 2004.
[25] P. Ruetschi, “Aging mechanisms and service life of lead-acid batteries,” Journal of
Power Sources, Vol. 127, pp. 33-44, 2004.
[26] J.H. Yan, W.S. Li, Q.Y. Zhan, “Failure mechanism of valve-regulated lead-acid
batteries under high-power cycling,” Journal of Power Sources, Vol. 133, pp. 135-140,
2004.
[27] R. Wagner, “Failure modes of valve-regulated lead/acid batteries in different
applications,” Journal of Power Sources, Vol. 53, pp. 153-162, 1995.
63
[28] A. Cooper, “Development of a lead-acid battery for a hybrid electric vehicle,”
Journal of Power Sources, Vol. 133, pp. 116-125, 2004.
[29] G.J. May, “Valve-regulated lead-acid batteries for stop-and-go applications,” Journal
of Power Sources, Vol. 133, pp. 110-115, 2004.
[30] P.T. Moseley, “High rate partial-state-of-charge operation of VRLA batteries,”
Journal of Power Sources, Vol. 127, pp. 27-32, 2004.
[31] A. Hande, “AC Battery Heating for Cold Climates,” Ph.D. Dissertation, The
University of Toledo, 2002.
[32] “Electric Double Layer Capacitor: BOOSTCAP Module Datasheet,” Maxwell
Technologies, www.maxwell.com.
[33] “ABC-150 Remote Operating System Manual, Revision 1.0,” Aerovironment Inc.,
August 1998.
[34] G. Stagg, A. El-Abiad, Computer Methods in Power System Analysis, McGraw
Hill, Inc., 1968.
[35] C. Lin, H. Peng, J. Grizzle, “Power Management Strategy for a Parallel Hybrid
Electric Truck,” Proceedings of the 2002 Mediterranean Control Conference, Lisbon,
Portugal, July 2002.
[36] Selmaware’s StampPlot Pro, www.selmaware.com
[37] “C505, C505C 8-bit CMOS Microcontroller User’s Manual 08.97,” Siemens,
August 1997.
Appendix I: System Software
See the enclosed CD for the software listings.
64
Appendix II: System Schematics
65
66
67
68
69