HIGH EFFICIENCY & HIGH PERFORMANCE CONVERTER/INVERTER SYSTEM
CONFIGURATIONS FOR HEV/EV TRACTION DRIVES

By
Craig B. Rogers

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Electrical Engineering
2012

ABSTRACT
HIGH EFFICIENCY & HIGH PERFORMANCE CONVERTER/INVERTER SYSTEM
CONFIGURATIONS FOR HEV/EV TRACTION DRIVES
By
Craig B. Rogers
Abstract—Hybrid and electric vehicles are becoming increasingly important as the cost
of available resources increase and as pollution and greenhouse gasses become more
pressing concerns. Hybrid and electric vehicles offer a means to increase efficiency and
decrease pollution and greenhouse gasses.

Whether the hybrid system is a series or series/parallel system, the vehicles discussed in
this work use an internal combustion engine (ICE), an AC electric generator, a
passive/active rectifier/inverter, a power storage system, a traction motor inverter, and a
traction motor/brake.

The value of the hybrid vehicle comes from its increased

efficiencies. The hybrid system decouples the ICE’s power and speed from that of the
vehicle. This decoupling is accomplished by replacing the traditional mechanical or
hydro mechanical transmission with power electronics and energy storage facilities. This
decoupling allows ICE operation at its most efficient power versus speed point. During
normal operation, the energy storage system can be used to handle the dynamic loads and
the internal combustion engine can handle the steady state loading. This hybrid system
also recaptures braking energy which is lost with conventional vehicle systems. The
power electronics facilitate the process via control of power flow, battery management,
motor control and generator control.

The existing hybrid vehicle power electronics topologies are discussed which include the
traditional back to back voltage source inverter to voltage source inverter configuration as
well as the newer back to back Z-Source type configurations. Each of these designs has
specific operation modes, control methods and features that are analyzed and discussed in
this document. The advantages and disadvantages of each are discussed and this will
form a basis upon which the new high efficiency designs proposed in this work can be
compared.

Finally, several new higher efficiency topologies are proposed and the basic features and
functionality of these proposed configurations are discussed. These new configurations
are then analyzed and quantified in terms of calculated efficiency.

Copyright by
CRAIG B. ROGERS
2012

ACKNOWLEDGEMENTS

I thank my family for their encouragement and the great value that they have contributed
to my life.

I thank my professors for their guidance direction and support throughout this process. I
especially thank Dr. Peng, Dr. Strangas, Dr. Mitra, and Dr. Zhu.

I also thank my colleagues, at Michigan State University’s Power Electronics and Motor
Drives Lab, for sharing their knowledge and friendship.

I thank the II-VI Foundation for their financial support. This work could not have been
done without their generosity.

v

TABLE OF CONTENTS
LIST OF FIGURES

VIII

KEYS TO SYMBOLS OR ABBREVIATIONS

XII

1. INTRODUCTION

1

1.2 Conventional Vehicles
1.3 Alternative Vehicles
1.4 Hybrid Vehicles
1.4.1 Series Hybrid Vehicle
1.4.2 Parallel Hybrid Vehicle
1.4.3 Series-Parallel Hybrid Vehicle
1.5 Fuel Cell Vehicles
1.6 Outline of Thesis

2
3
5
7
8
10
12
13

2. REVIEW OF EXISTING HYBRID ELECTRIC VEHICLE
POWER ELECTRONICS TOPOLOGIES AND THEIR
OPERATING PRINCIPLES
17
3. EVALUATION OF MAJOR HEV COMPONENTS AND
OVERALL OPERATION OF HEV
29
3.1 Characteristics of the Internal Combustion Engine (ICE)
3.2 Characteristics of the Hybrid Battery Pack
3.2.1 Comparison of the high voltage battery vs. the low voltage battery. High
voltage battery option.
3.2.2 Comparison of the high voltage battery vs. the low voltage battery. Low
voltage battery with DC/DC converter option.
3.2.3 Summary of High Voltage Battery Option vs. Low Voltage Battery Option
3.3 Characteristics of Inverters and Converters
3.4 Traction Motor Operating Conditions
3.5 Summary of Hybrid Major Components

31
33
34
37
38
38
41
43

4.
PRESENTATION AND BRIEF ANALYSIS OF
PROPOSED TOPOLOGIES
46
4.1 VSI / VSI with Conventional Mid-Link DC / DC Converter
4.2 VSI / VSI with “H” Bridge Coupled Battery Pack
4.3 VSI / VSI Structure, LV Battery with Boost Converter, & DC-Link with Buck
Converter
4.4 VSI / VSI Structure, LV Battery with Buck-Boost Converter, & DC-Link with
Buck Converter
vi

47
49
51
52

4.5 VSI / VSI Structure, LV Battery, DC-Link with Integrated Magnetic Boost
Converter Using Open Circuit Zero States of Inverter Phase Legs
4.6 VSI / VSI Structure, LV Battery, DC-Link with Bi-Directional Buck and Boost
Converter
4.7 Back to Back Current Fed Quasi-Z-Source Inverters, LV Battery
4.8 Back to Back Voltage Fed Quasi-Z-Source Inverters, LV Battery
4.9 Summary of proposed topologies

5. LOSS ANALYSIS OF SELECTED TOPOLOGIES

54
59
61
62
64

65

5.1 Base Line:
65
5.1.1 Evaluation of Efficiency of the topologies.
65
5.1.2 Experimental Results
71
5.1.3 Base Line Continued
79
5.2 Proposed Topologies
82
5.2.1 VSI / VSI with Conventional Mid-Link DC / DC Boost Converters
83
5.2.2 VSI / VSI Structure, LV Battery with Boost Converter, & DC-Link with Buck
Converter
85
5.2.3 VSI / VSI Structure, LV Battery with Buck-Boost Converter, & DC-Link with
Buck Converter
87
5.2.4 VSI / VSI Structure, LV Battery, DC-Link with Bi-Directional Buck and Boost
Converters
88
5.2.5 Back to Back Voltage Fed Quasi-Z-Source Inverters, LV Battery
89
5.2.6 Back to Back Current Fed Quasi-Z-Source Inverter, LV Battery
91
5.3 Conclusion
104

6.

FUTURE WORK

110

REFERENCES

112

vii

LIST OF FIGURES
FIGURE 1 TYPICAL INTERNAL COMBUSTION ENGINE (ICE) VEHICLE .................................. 3
FIGURE 2 TYPICAL ELECTRIC VEHICLE (EV) ....................................................................... 4
FIGURE 3 TYPICAL SERIES HEV .......................................................................................... 7
FIGURE 4 TYPICAL PARALLEL HEV ................................................................................... 10
FIGURE 5 TYPICAL SERIES-PARALLEL HEV [5] ................................................................. 11
FIGURE 6 TYPICAL SERIES-PARALLEL HEV WITH PLANETARY GEAR SET [5] ................... 12
FIGURE 7 TYPICAL FUEL CELL VEHICLE ............................................................................ 13
FIGURE 8 TRADITIONAL INPUT RECTIFIER/VSI TO OUTPUT VSI. ....................................... 21
FIGURE 9 TRADITIONAL INPUT RECTIFIER/VSI TO OUTPUT VSI WITH............................... 22
FIGURE 10 INPUT RECTIFIER/VSI TO OUTPUT CSI WITH Z-SOURCE IMPEDANCE ............... 24
FIGURE 11 INPUT RECTIFIER/CSI TO OUTPUT VSI WITH Z-SOURCE IMPEDANCE ............... 26
FIGURE 12 INPUT RECTIFIER/VSI TO OUTPUT CSI WITH Z-SOURCE IMPEDANCE ............... 28
FIGURE 13 SERIES HYBRID ELECTRIC CONFIGURATION ..................................................... 30
FIGURE 14 TYPICAL TORQUE VS. SPEED CURVES OF A DIESEL ICE.................................... 32
FIGURE 15 TRADITIONAL BACK TO BACK VSI / VSI HYBRID TOPOLOGY ........................... 33
FIGURE 16 SERIES COMBINATION OF BATTERY CELLS ..................................................... 36
FIGURE 17 IGBT MODULE USED IN INVERTERS AND CONVERTERS .................................... 39
FIGURE 18 EPA URBAN DYNAMOMETER DRIVING SCHEDULE (UDDS) ............................ 42
FIGURE 19 TRADITIONAL BACK TO BACK VSI / VSI HYBRID TOPOLOGY ........................... 43
FIGURE 20 VSI / VSI WITH CONVENTIONAL MID-LINK DC / DC CONVERTER .................. 47
FIGURE 21 VSI / VSI WITH CONVENTIONAL MID-LINK DC / DC CONVERTER. ................. 48
FIGURE 22 VSI TO VSI TOPOLOGY WITH AN “H” BRIDGE COUPLING THE BATTERY ........... 49
viii

FIGURE 23 VSI TO VSI TOPOLOGY WITH AN “H” BRIDGE COUPLING THE BATTERY PACK TO
THE DC-LINK. ............................................................................................................ 50
FIGURE 24 VSI / VSI STRUCTURE, LV BATTERY WITH BOOST CONVERTER, & DC-LINK
WITH BUCK CONVERTER ............................................................................................ 51
FIGURE 25 VSI / VSI STRUCTURE, LV BATTERY WITH BOOST CONVERTER, & DC-LINK
WITH BUCK CONVERTER. ........................................................................................... 52
FIGURE 26 VSI / VSI STRUCTURE, LV BATTERY WITH BUCK-BOOST CONVERTER, & DCLINK WITH BUCK CONVERTER ................................................................................... 52
FIGURE 27 VSI / VSI STRUCTURE, LV BATTERY WITH BUCK-BOOST CONVERTER, & DCLINK WITH BUCK CONVERTER. .................................................................................. 53
FIGURE 28 VSI / VSI STRUCTURE, LV BATTERY, DC-LINK WITH INTEGRATED MAGNETIC
BOOST CONVERTER ................................................................................................... 54
FIGURE 29 VSI / VSI STRUCTURE, LV BATTERY, DC-LINK WITH INTEGRATED MAGNETIC
BOOST CONVERTER ................................................................................................... 56
FIGURE 30 INVERTER SWITCHING STATE, DC-LINK VOLTAGE, AND BACK EMF ................ 57
FIGURE 31. SWITCH STATES AND DC-LINK VOLTAGE. ....................................................... 58
FIGURE 32 VSI / VSI STRUCTURE, LV BATTERY, DC-LINK WITH BI-DIRECTIONAL BUCK 59
FIGURE 33 VSI / VSI STRUCTURE, LV BATTERY, DC-LINK WITH BI-DIRECTIONAL BUCK 60
FIGURE 34 BACK TO BACK CURRENT FED QUASI-Z-SOURCE INVERTERS, LV BATTERY .. 61
FIGURE 35 BACK TO BACK CURRENT FED QUASI-Z-SOURCE INVERTERS, LV BATTERY. . 61
FIGURE 36 BACK TO BACK VOLTAGE FED QUASI-Z-SOURCE INVERTERS, LV BATTERY .. 62
FIGURE 37 BACK TO BACK VOLTAGE FED QUASI-Z-SOURCE INVERTERS, LV BATTERY. . 63
FIGURE 38. VOLTAGE AND CURRENT WAVEFORMS OF A TYPICAL SEMICONDUCTOR
SWITCHING DEVICE .................................................................................................... 67
FIGURE 39. EXPERIMENTAL SET UP SCHEMATIC DIAGRAM ................................................. 72
FIGURE 40 EXPERIMENTAL SETUP ..................................................................................... 73
FIGURE 41 AMPLITUDE MODULATED ROTOR POSITION SIGNALS. ....................................... 74

ix

FIGURE 42 AMPLITUDE MODULATED ROTOR POSITION SIGNALS ........................................ 75
FIGURE 43 CHASSIS DYNO. USED TO OBTAIN EXPERIMENTAL RESULTS.............................. 76
FIGURE 44 EXPERIMENTAL SETUP..................................................................................... 77
FIGURE 45 MORE OF THE EXPERIMENTAL SETUP. ............................................................... 78
FIGURE 46 EXPERIMENTAL RESULTS SHOWING FOUR OPERATING POINTS .......................... 78
FIGURE 47. POWER LOSS CURVES FOR THE TRADITIONAL BACK TO BACK VSI/VSI............ 82
FIGURE 48 BACK TO BACK VSI TO VSI WITH BOOST CONVERTERS IN THE DC-LINK ......... 83
FIGURE 49 LOSS GRAPHS (A, B, AND C) FOR THE BACK TO BACK VSI TO VSI WITH BOOST
CONVERTERS IN THE DC-LINK ................................................................................... 83
FIGURE 50 BACK TO BACK VSI / VSI TOPOLOGY WITH BATTERY BOOST CONVERTER AND TI
DC-LINK BUCK CONVERTER. ..................................................................................... 85
FIGURE 51 LOSS GRAPHS (A, B AND C) FOR THE BACK TO BACK VSI / VSI TOPOLOGY ...... 85
FIGURE 52 BACK TO BACK VSI / VSI TOPOLOGY WITH BATTERY BUCK-BOOST CONVERTER
................................................................................................................................... 87
FIGURE 53 LOSS GRAPHS UNDER SCENARIO ONE CONDITIONS (GENERATOR AT 50KW) ..... 87
FIGURE 54 BACK TO BACK VSI / VSI TOPOLOGY WITH GI BI-DIRECTIONAL BUCK-BOOST . 88
FIGURE 55 LOSS GRAPHS UNDER SCENARIO ONE CONDITIONS (GENERATOR AT 50KW) ..... 88
FIGURE 56 BACK TO BACK VOLTAGE FED QUASI Z SOURCE INVERTER TOPOLOGY. ............ 89
FIGURE 57 LOSS GRAPHS (A, B, AND C) FOR THE BACK TO BACK VOLTAGE FED QUASI Z
SOURCE INVERTER TOPOLOGY .................................................................................... 89
FIGURE 58 BACK TO BACK CURRENT FED QUASI Z SOURCE INVERTER TOPOLOGY. ............ 91
FIGURE 59. POWER LOSS COMPARISONS BETWEEN THE TRADITIONAL TOPOLOGY AND ...... 91
FIGURE 60. OPERATIONAL STATES (A, B, AND C) OF THE CF-QZSI [27] ............................. 94
FIGURE 61. VOUT/VIN VS. ACTIVE DUTY CYCLE. [27] ........................................................ 96
FIGURE 62. GENERATOR END OF THE BACK TO BACK CF-QZSI TOPOLOGY ........................ 98

x

FIGURE 63. (A) AND (B) GI VOLTAGE IS BEING BUCKED FROM THE GI LEVEL .................... 99
FIGURE 64. THE SWITCHING STATES USED TO START THE DIESEL ICE .............................. 100
FIGURE 65 COMPARATIVE POWER LOSS GRAPHS FOR THE TRADITIONAL, ........................ 106

xi

KEYS TO SYMBOLS OR ABBREVIATIONS
ac voltage at the generator/starter motor - Line to Line (VG-LL)

ˆ

ac voltage at the generator/starter motor - peak ( VG )
ac voltage at the generator/starter motor - RMS ( VG )
ac voltage at the traction motor - Line to Line (VTM-LL)

ˆ

ac voltage at the traction motor - peak ( VTM )
ac voltage at the traction motor - RMS ( VTM )
Acceleration of automobile (aa)
Acceleration of bus (ab)
amperes (A)
Angular velocity (ω)
Average Line Current (IAV or Iav)
Battery terminal voltage (VBAT)
Battery Voltage (VBATT)
Boost factor (B)
Buck–boost factor (BB)
Capacitor voltage (VC)
Collector current (IC)

xii

Conditional probability of a short circuit failure (p)
Conditional probability of an open circuit failure (q)
Constant power speed ratio (CPSR)
Corporate Average Fuel Economy (CAFE)
Cumulative distribution function of time to failure for a reliability study (F(t))
Current Fed Z-Source Inverter (CF-ZSI)
Current Source Inverter (CSI)
DC bus voltage (Vcc)
DC voltage (Vdc)
DC-Link voltage on the Generator Inverter End (VDC_GI)
DC-Link voltage on the Traction Inverter End (VDC_TI)
Diode conduction loss (Pd)
Duty cycle of switching device (D)
Electric vehicle (EV)
Electro Magnetic Interference (EMI)
Environmental Protection Agency (EPA)
Filter voltage (VF)
Force of acceleration of a bus (Fb)
Force of acceleration of an automobile (Fa)
Fuel cell (FC)

xiii

Fuel cell hybrid electric vehicle (FC HEV)
Fuel cell vehicle (FCV)
Generator Inverter (GI)
Generator Power, Three Phase (P3ϕ )
Generator Voltage (VG)
Hybrid electric vehicle (HEV)
Hybrid vehicle (HV)
IGBT conduction loss (Pi)
Insulated Gate Bi-Polar Junction Transistor Module (IGBT)
Internal combustion engine (ICE)
Kilowatt (kW)
Kilowatt hour (kWhr)
Line to line voltage (VLL)
Mass of automobile (ma)
Mass of buss (mb)
Maximum collector current (IcM)
Modulation index (M)
Modulation index (M)
Newton Meter (Nm)
number of components in series in a reliability study (n)
number of parallel paths in a reliability study (m)
xiv

Open Circuit Zero Switching Period (TOCZ) or (T1)
Open Circuit Zero switching state (OCZ)

ˆ
Peak Line Current ( I )
Phase A Lower switch gate control signal (San)
Phase A upper switch gate control signal (Sap)
Phase A voltage reference (Va*)
Phase angle between phase voltage and line current (Ө)
Phase B Lower switch gate control signal (Sbn)
Phase B upper switch gate control signal (Sbp)
Phase B voltage reference (Vb*)
Phase C Lower switch gate control signal (Scn)

Phase C upper switch gate control signal (Scp)
Phase C voltage reference (Vc*)
Pulse width modulation (PWM)
Rated collector current (IcN)
Rated collector to emitter voltage (VceN)
Rated diode forward voltage (VFN)
Rated fall time (tfN)

xv

Rated recovery charge (Qrrn)
Rated recovery time (trrN)
Rated rise time (trN)
Reliability as a function of time (R(t))
Reverse Blocking Insulated Gate Bi-Polar Junction Transistor Module (RB-IGBT)
Reverse recovery loss of the diode (Prr)
Revolutions Per Minute (rpm)
Series Hybrid Electric Vehicle (SHEV)
Shoot-through duty cycle (D0)
Short Circuit Zero Switching Period (TSCZ)
Sinusoidal pulse width modulation (SPWM)
State of charge (SOC)
Switching cycle interval (T)
Switching Device Power (SDP)
Switching Device Power Ratio (SDPR)
Switching frequency (Fs)
Threshold collector emitter voltage (Vco)
Threshold diode forward voltage (VFo)

xvi

Time during shoot through (T0)
Torque (τ)
Traction Inverter (TI)
Traction Motor Power, Three Phase (P3ϕ_TM)
Traction Motor Voltage (Vm)
Turn off loss of the IGBT (Poff)
Turn on loss of the IGBT (Pon)
Upper shoot-through envelope (Vp)
Upper shoot-through envelope reference signal (Vpref)
Urban Dynamometer Driving Schedule (UDDS)
Voltage across filter (VF)
Voltage Fed Z-Source Inverter (VF-ZSI)
Voltage gain (MB)
Voltage line to line (VLL)
Voltage of battery (VBATT) or (Vbatt)
Voltage Source Inverter (VSI)
Wide open or full throtle (WOT)

xvii

Zero emission vehicle (ZEV)
Z-Source Inductor Current (ILZ)
Z-Source Inverter (ZSI)

xviii

1. INTRODUCTION
Alternatively powered vehicles are becoming of great importance. Several factors are
causing the need for change from traditional vehicles. As the population of the planet
continues to grow, and as the industrialization of these populations continue to grow, the
rate of depletion of traditional fuel sources (fossil fuels – oil, and coal primarily) is also
increasing. Along with this increased global demand for power comes the increased
pollution from consuming fuels. Both the decreasing supply of fossil fuels and the
increasing level of pollutants in our environment are becoming very significant concerns.
As a result, there is a growing need to search for ways in which these problems may be
reduced or solved. Both alternative fuels and ways to increase efficiency in the use of
fuels are being sought after. One such approach to increase efficiency is the development
of hybrid and electrical vehicles which are much more fuel efficient and which produce a
smaller amount of pollutants. This introduction will provide a discussion of traditional
vehicles, electric vehicles, fuel cell vehicles, and hybrid vehicles.

Because an

understanding of the hybrid electric vehicle and electric vehicle is necessary to
understand how the higher efficiency topologies achieve their greater efficiency
performance, a brief discussion of hybrid and electric vehicle types follows.

1

1.2 Conventional Vehicles

The tradition of fuel powered vehicles started in 1769 with a self-propelled military
tractor invented by the French engineer and mechanic, Nicolas Joseph Cugnot (17251804). Much before this date, conceptual designs had been made by both Leonardo Da
Vinci and Isaac Newton.[1]

In 1823, Samuel Brown patented the first internal

combustion engine to be applied industrially. It was compression-less and based on the
"Leonardo cycle,"

In 1838, a patent was granted to William Barnet (English) for an

engine using in-cylinder compression. In 1854, the Italians Eugenio Barsanti and Felice
Matteucci patented the first working efficient internal combustion engine in London but
did not go into production with it. It was similar in concept to the successful Otto Langen
indirect engine, but not so well worked out in detail[2]. Various fuels were used over
these years. By far the most prevalent designs today use gasoline or diesel fuel. The
typical road vehicle in use today is based upon a standard platform. That is a gasoline or
diesel fueled internal combustion compression engine. The engine is coupled to the
remainder of the drive train through a multiple gear ratio transmission.

From the

transmission, power is delivered to the drive wheels through a drive shaft/s and a
differential gear box. The engines are of the cylindrical compression type, and the
transmission usually has three or more discrete gear ratios.

The function of the

differential gear box is to allow one of the drive wheels to rotate at a different speed than
the other drive wheel while both are delivering power to the road. Each of these drive
train components is mechanical. See Figure 1 for a schematic of the typical conventional
vehicle.

2

ICE

Mechanical
Power Flow

Transmission

Wheel

Differential
Gears

Wheel

Figure 1 Typical Internal Combustion Engine (ICE) Vehicle

Hybrid vehicles use some combination of these mechanical components as well as
electrical components to drive the vehicle.

1.3 Alternative Vehicles

Various countries and governments across the globe are trying to foster the development
of alternative vehicles that mitigate some of these concerns. Alternatives are necessary to
eventually create vehicles that do not further deplete crude oil reserves and that produce
pollutants at a much lower rate. The population of the world continues to grow, more
nations are becoming industrialized, and burning of fossil fuels is believed to contribute
to global warming. There are many reasons to turn to alternatives that would alleviate
some or all of these concerns. One such alternative is the electric vehicle (EV). This
vehicle does not have an internal combustion engine and does not burn any fuels on
3

board. Its power is usually received from the electric utility and stored in batteries that
are on board the vehicle.

Battery

Inverter

Wheel

Mechanical
Power Flow
Electrical
Power Flow

Motor /
Generator

Wheel

Figure 2 Typical Electric Vehicle (EV)

The stored power is then applied to eclectic motors to drive the vehicle rather than using
an ICE. This type of vehicle is very efficient as compared to conventional vehicles.
Presently, the fuel cost per mile (cost per kilo-watt hour (kWhr)) is about 3 cents per mile
as compared to 14 cents per mile for gasoline in the conventional automobile. One of the
main advantages of the EV is that it is a zero emission vehicle (ZEV). Although there is
some pollution created when the electricity is generated at the utility power plant, this is
insignificant compared to emissions from conventional ICE vehicles. EVs are also very
efficient when compared to conventional ICE vehicles; only about 20% of the chemical

4

energy in gasoline is converted into work at the wheels of an ICE vehicle, 75% or more
of the energy from a battery reaches the wheels of an EV [3]. EVs also provide strong
performance, due to the fact that electric motors provide nearly peak torque even at low
speeds. Cost, performance, and convenience of an EV all depend on the battery. There
are several types of batteries commonly used in EVs, advanced lead-acid batteries,
nickel-metal hydride, and lithium polymer batteries [3].The drawback to this type of
vehicle is that it takes more time to recharge the batteries and recharging stations are not
yet commonly available. Because of the limited storage capacity of the batteries, the
range of these vehicles is also more limited.

1.4 Hybrid Vehicles

Hybrid vehicles (HV) are growing in popularity for their fuel efficiency and lower
emissions. Fuel costs are increasing and the world’s fuel reserves are continuing to
decrease. Since the world’s supply of crude oil is based in just a few countries, a nation’s
dependence on oil also makes it dependent upon the oil producing nation. Also, pollution
issues are increasing in weight. These are the main reasons why the demand for HVs is
increasing and why governments are encouraging the sales and development of these
vehicles.

A HV is any vehicle that combines two or more sources of power to propel the
vehicle. Some of the first HVs were busses, trains, and submarines. Hybrid methods are
now being applied to passenger cars, small trucks, and industrial vehicles such as
excavators and busses. There are three main configurations of HVs. They are the Series
Hybrid, Parallel Hybrid, and Series/Parallel Hybrid. They are summarized below.
5

A common type of hybrid electric vehicle (HEV) combines the energy storage of the EV
with an energy source such as an ICE. HEVs have higher efficiency levels due to
regenerative braking, and due to their ability to control power from the ICE.

By

controlling the power flow, it is possible to utilize the most efficient points of operation.
Regenerative braking takes advantage of the concept that energy can neither be created
nor destroyed – it just changes form. The conventional vehicle converts the kinetic
energy of the moving vehicle into heat energy when the brakes are applied to slow the
vehicle. This radiated heat is lost to the atmosphere. The HEV converts this inertia into
electrical energy by using the traction motor as an electric generator. The production of
electrical energy by the generator also produces a resistive torque against the free rotation
of the drive wheels which slows the vehicle. This electrical energy is stored as potential
energy in the vehicle’s battery bank and or ultra-capacitors. In this way, energy that
would have been wasted is saved for later use. Also, the HEV can use a smaller ICE.
The ICE only needs to supply the consistent steady state load while the battery and
electric motor can provide power for the dynamic load. Some examples of the dynamic
load would be the additional power needed to accelerate and to climb hills. The smaller
ICE uses less fuel that the larger engines do. The smaller engine requires less fuel into
the smaller cylinders to achieve the correct air to fuel ratios and expends less energy
pumping the air and combustion gases in and out of the cylinders. Also, an ICE has
moving parts which involve friction. The smaller parts have less surface fraction. Also,
the engines have linear reciprocating motion components which must undergo four
accelerations for each crank shaft revolution. Obviously, the lower mass components
require less energy.

6

1.4.1 Series Hybrid Vehicle
The series HEV delivers all of the power to the drive wheels via electric power. None of
the power is delivered with direct mechanical connections.

For example, a typical

configuration uses an ICE connected to a generator. The generator is connected to the
traction motor through an electric power inverter and conductors. The power inverter is
usually connected to power storage devices such as batteries and or ultra-capacitors. See
Figure 3.

ICE
Battery
Generator
Power
Electronics

Wheel

Motor /
Generator

Mechanical
Power Flow
Electrical
Power Flow

Wheel

Figure 3 Typical Series HEV
The series HEV uses an ICE which runs for extended periods of time at
maximum efficiency, and can shut off when the batteries reach a selected state of charge
(SOC). The ICE is then allowed to operate at the average required power level and the

7

batteries will accept power when this average exceeds instantaneous traction power
demand. The battery likewise provides power when the traction power required exceeds
the ICE’s average power provided.
In the series HEV configuration the electric motors and power electronics must be
sized for the whole load whereas the parallel case can use a smaller inverter and motor.
Still, the ICE is smaller than that in the conventional vehicle and is comparable to that of
the parallel case. Also, this design does not require the additional equipment needed for
the parallel case. In higher power applications (busses, freight trains, subway trains), this
is a very effective design application.
The series HEV suffers due to the number of energy conversions. The engine
converts petroleum potential energy to kinetic energy (rotational motion); the generator
converts this kinetic energy to electrical energy. The electrical energy is then converted
to chemical energy in the battery and to mechanical (kinetic energy) in the traction motor.
However, the series HEV does not require a lossy hydro mechanical transmission. Also,
batteries have limited life expectancy.

New battery technology has made great

improvements with this and there are several batteries that have long life periods when
the battery state of charge (SOC) is maintained within a fairly narrow band and when the
charge and discharge rates are maintained within specified limits [3].

1.4.2 Parallel Hybrid Vehicle
The parallel hybrid vehicle has two power flow paths. One is mechanical from the ICE
to the wheels, and the other is electrical from the generator and batteries to the electric
8

motor and then mechanically to the wheels. The battery is recharged from regenerative
braking and by operating the ICE at a power level in excess of the instantaneous traction
power demand.

This excess power is delivered to the batteries through a gearing

mechanism, to a motor (being operated as a generator), to the power inverter/converter
and then to the battery. The main advantages of this system are that both the engine and
the motor size can be minimized since both can deliver power to the wheels at the same
time to accommodate the peak load requirements. Also, there are potentially less energy
conversions for the mechanical power flow path. On the other hand this system typically
utilizes a hydro mechanical transmission which has losses, and the control is more
complicated than the series HEV [3].

9

Mechanical
Power Flow
Battery
Electrical
Power Flow
Power
Electronics
ICE

Motor /
Generator

Transmission

Wheel

Gear Box

Wheel

Figure 4 Typical Parallel HEV

1.4.3 Series-Parallel Hybrid Vehicle

The series-parallel HEV combines features from both of the preceding cases. This
system can operate in series mechanical mode, series electrical mode, or any
combined ratio of the two. As a result, the control system can choose the most
advantageous operation mode for the given circumstances.
Figure 5 shows a power flow diagram for a series-parallel hybrid. Also notice the
10

planetary gear component in this figure. This is also a very significant part of the
efficiency of the design. The planetary gears (see Figure 6 for a diagram) deliver power
from the ICE to the wheels or to the generator. It also couples power transfer to or from
the generator. The generator is an instrumental part of this control. The generator is used
to create either positive or negative torque so that the desired gear ratio between the ICE
and vehicle wheels is achieved.

As a result of the ideal gear ratio, the ICE is

continuously operated at its most efficient point for the given power demand. Also, the
planetary gear system does not use lossy hydraulic torque converters as the conventional
vehicle does.

τrωr

τeωe

fuel
engine

planetary

Pout=τsωs
counter
shaft

τgωg

To wheels

τmωm

generator

motor
+

battery
Figure 5 Typical Series-Parallel HEV [5]

11

4 quadrant motor/
generator

Planetary Gear Set

Ring gear

τeng. ωeng.

ω

τ
Sun gear

ω

τ

N3
N4

Ring gear N2

ω

ω

τ

4 quadrant motor/
generator

τ

N1

4 quadrant
torque and speed
to differential
gear set

N5

Figure 6 Typical Series-Parallel HEV with Planetary Gear Set [5]
This system benefits from the reduced size ICE, generator, power inverter, and traction
motor due to its parallel functionality. Also, the operating system is able to select the
most advantageous operation mode for the given circumstances.

1.5 Fuel Cell Vehicles
Fuel cell vehicles (FCV) are still being developed and have not yet achieved commercial
production. However, this type of vehicle offers a great deal of benefits in terms of
efficiency, and very low emissions. Its environmental impact is similar to that of an EV,
yet the FCV has a large range comparable to the conventional EV. The FCV utilizes
fuels such as hydrogen, methanol, or gasoline to produce electrical power without
combustion. The waist product of the reaction is typically H20 and CO2. The drive train
of a FCV thus becomes: Fuel Cell, Inverter, and Traction Motor. Most designs will likely
12

include a battery as well. The typical fuel cell does not have the ability to respond
quickly to power demand changes. Therefore, the fuel cell can be used to supply average
power and the battery can be used to supply or absorb the difference between average
power and instantaneous power. Adding regenerative braking to the FCV makes the
design very fuel efficient. Also, the infrastructure for these fuels does not yet exist.
Below is a diagram of the FCV structure.

Battery

Mechanical
Power Flow

Fuel Cell
Power
Electronics

Wheel

Motor /
Generator

Electrical
Power Flow

Wheel

Figure 7 Typical Fuel Cell Vehicle

1.6 Outline of Thesis
The goal of this research is to review the operational characteristics of the major
components of the HEV, investigate the conditions under which they operate as well as
the overall operating conditions of the overall vehicle with the hope of finding
13

operational modes or systems that could then be devised to achieve greater efficiency
than presently existing technology. The traction drive system will deliver power to and
from the ICE via a generator/starter motor and it will deliver power to and from the
traction motor/generator. This system will also include interface and control of storage
batteries.

Chapter 2: Reviews the current HEV power electronics technologies, and their operating
principles. The hybrid operational modes are briefly discussed. These are the mode of
operation that a hybrid vehicle must be able to achieve in order to be considered
effective. Finally, five existing hybrid vehicle topologies are shown.

Chapter 3: Discusses the major components of the hybrid vehicle system. Specifically,
these are the internal combustion engine, permanent magnet alternating current motors /
generators, battery packs, DC to AC inverters, DC to DC converter, and the vehicle itself.
It is the characteristics of these components that present operating constraints of which
the new proposed solutions must abide. However, it is these same constraints that are
herein exploited to achieve new and higher efficiency topologies.
3.1 Internal Combustion Engine
3.2 Hybrid Battery Pack
3.3 Inverters and Converters
3.4 Traction Motor and vehicle operating conditions
3.5 Summary of characteristics and presentation of the generator inverter to
traction inverter decoupled voltage concept.

14

Chapter 4: Presents several new topologies the purpose of which is to employ voltage
decoupling and thereby reap greater efficiencies. Each topology to be studied in this
work is presented in this section along with a brief illustration of power flow through the
topology.
4.1 VSI / VSI with mid-link conventional DC/DC converter.
4.2 VSI / VSI with “H” bridge coupled battery stack.
4.3 VSI / VSI Structure, LV Battery with Boost Converter, & DC-Link with Buck
Converter.
4.4 VSI / VSI Structure, LV Battery with Buck-Boost Converter, & DC-Link with
Buck Converter.
4.5 VSI / VSI Structure, LV Battery, DC-Link with Integrated Magnetic Boost
Converter Using Open Circuit Zero States of Inverter Phase Legs.
4.6 VSI / VSI Structure, LV Battery, DC-Link with Bi-Directional Buck and
Boost Converter.
4.7 Back to Back Current Fed Quasi-Z-Source Inverters, LV Battery.
4.8 Back to Back Voltage Fed Quasi-Z-Source Inverters, LV Battery.
4.9 Summary of proposed topologies.

Chapter 5: Presents the above topologies which were functional and analyzes their
power losses comparing them to the traditional topology.

Before this can be done

however, the conditions, components, and operating conditions of the analysis are
described. Experimental results are presented to confirm the voltage decoupling concept

15

and to verify the accuracy of the calculations.
5.1 Development of the base line.
5.2 Proposed Topologies
5.3 Conclusions

Chapter 6: The work and results presented in this writing have uncovered yet additional
ways in which hybrid vehicles can be further improved. Chapter six suggests some
future work toward this extent that can begin where this work ends.

16

2. REVIEW OF EXISTING HYBRID ELECTRIC VEHICLE
POWER ELECTRONICS TOPOLOGIES AND THEIR
OPERATING PRINCIPLES
The purpose of this thesis is to investigate traction drive systems for a series hybrid
electric bus. The purpose of the power electronics is to control the flow of power from or
to the ICE’s generator, to or from the batteries, and to or from the traction motor. The
voltage must be managed in terms of ac frequency and magnitude. The amount of power
flow to or from the system batteries must be managed to control the state of charge of the
batteries. Power flow to or from the traction motor must be controlled to provide the
desired acceleration/deceleration and speed of the vehicle. In this process, of controlling
the power flow for the battery, traction motor, and generator, all should be managed so
that optimal efficiency is achieved. The existing topologies are shown here and will form
a base against which the newly proposed topologies of this work can be compared.
Below is a summary of the existing topologies which will be discussed in more detail.



Traditional Input Rectifier/VSI to Output VSI



Input Rectifier/VSI to Output CSI with Z-Source impedance network
(battery in parallel with Z’s capacitor)



Input Rectifier/CSI to Output VSI with Z-Source impedance network
(battery in parallel with Z’s capacitor)



Input Rectifier/VSI to Output CSI with Z-Source impedance network
(battery in parallel with Input Converter)

Each topology and its operating principles will be briefly reviewed. The traditional Input
Rectifier/VSI to Output VSI will be discussed first due to its historical familiarity. This
17

topology will be used to discuss the general function of the power electronics to deliver
and control power.

The hybrid vehicle has the flexibility of choosing where power is to come from and
where it is to be delivered to. The goal is to make choices that optimize efficiency. In
general, this means that the ICE and generator are to be operated in a narrow range of
angular velocity and a relatively small region of torque to supply the average power
required. The batteries are to provide the “dynamic” load. That is the power above and
below the average power needed. In a vehicle route, the vehicle must accelerate and
decelerate many times. The total energy expended by the vehicle on its route divided by
the amount of time it takes, gives us the average power. In this example, while the
vehicle is accelerating, the traction motors typically require more power than this average
power level.

This additional momentary power variance is to be provided by the

batteries. With this method, the ICE power vs. angular velocity is chosen to be its most
efficient operating point. Operation at the most efficient point is not possible with the
conventional vehicle and its transmission since the ICE alone must provide all of the
range of angular velocity and of power that the vehicle’s wheels/transmission require.
Also, the hybrid vehicle is capable of recapturing deceleration energy which is wasted by
conventional vehicles. To move power consistent with these goals, several operation
modes are encountered. These modes are summarized here since it is desirable for any
new topology to be able to achieve each mode.

18

Hybrid Operational Modes:
1. Battery and Traction Motor mode
The traction motor is lightly loaded, the battery is supplying power
to the traction motor, and the generator is off.
2. Battery, Starter Motor and Traction Motor mode
The battery is supplying power to the traction motor and
simultaneously providing power to the generator. The generator is
momentarily being used as a motor to start the internal combustion
engine (ICE).
3. Generator, Battery Discharge and Traction Motor mode
The traction motor is medium to heavily loaded. The generator
and the battery are both supplying power.
4. Generator, Battery Charging and Traction Motor mode
During this mode, the Traction Motor load has decreased and the
battery transitions from discharging to charging.
5. Regenerative Braking and Battery Charging mode
During this mode, the load at the traction motor becomes
negative and the traction motor functions as a regenerative
brake (generator). The battery is charging.

Existing Topologies
The first two schematics are of the traditional hybrid bus back to back voltage source

19

inverter (VSI) topologies and the last three are “Z” source based topologies. Recall that
the VSI must have a voltage source that is greater than the desired output voltage level.

VLL 
Where:

3 Vdc
M
2 2

VLL is the line to line RMS output voltage
Vdc is the dc input voltage
M is the modulation index [ 0  M  1.0 ]

Therefore, the VSI is considered a buck converter. If Vdc cannot be as high as is needed,
than it would be necessary for a DC/DC converter to be added to the circuit architecture
to achieve the required input voltage. This is one of the limitations of the VSI topologies
that are overcome with the family of Z-Source topologies.

20

Generator
Inverter
Sas-p Sbs-p Scs-p

Generator
/ Starter
Motor Rs

Traction Inverter

S1T-p

S2T-p

S3T-p
Traction
Motor / Regenerative
Brake

C1

Ls

as

LT

Vd

RT

bs
VG

Vm

cs

Sas-n Sbs-n Scs-n

S1T-n

S2T-n

S3T-n

Figure 8 Traditional Input Rectifier/VSI to Output VSI.
The left end of the schematic shows the generator which is connected to the ICE. This end is considered the power flow input. The
term rectifier is also used since the input converter may be used as a passive rectifier. Voltage sources in the center and a VSI at the
output

21

Generator
Inverter
Sas-p Sbs-p Scs-p
Generator /
Starter
Motor Rs Ls

LVBatt

as
bs

VG

Traction Inverter
S1T-p S2T-p S3T-p

C1
DC-DC
Converter

LT R T

Traction
Motor / Regenerative
Brake

Vd
Vm

cs

Sas-n Sbs-n Scs-n

S1T-n

S2T-n

S3T-n

Figure 9 Traditional Input Rectifier/VSI to Output VSI with
DC/DC Converter.
The left end of the schematic shows the generator which is connected to the ICE. This end is considered the power flow input. The
term rectifier is also used since the input converter may be used as a passive rectifier. Battery and DC/DC converter in the center and
a VSI at the output

22

The next circuit architecture is the Input Rectifier/VSI to Output CSI with Z-Source
impedance network and the battery is in parallel with Z’s capacitor. Because of the
bidirectional current flow allowed by the input converter (on the left of the schematic),
the input converter is called a VSI. Notice that the output converter (right side) only
allows one directional current flow but bidirectional voltage blocking. As such, it is
called a CSI. The network in the center is made up of the unique combination of
capacitors

and

inductors

and

is

called

23

the

Z-Source

(Impedance-Source).

Generator Inverter
Sas-p Sbs-p Scs-p

Generator /
Rs Ls
Starter
Motor

VG

+
VB1
-

as

Traction Inverter
+ L1 -

+ +
C1 C2
- -

bs

S1T-p S2T-p

+
VB2
-

S3T-p
Traction
Motor / ReLT RT generative
Brake

+

Vi
-

cs

Sas-n Sbs-n Scs-n

- L2 +

Vm

S1T-n S2T-n S3T-n

Figure 10 Input Rectifier/VSI to Output CSI with Z-Source impedance
Network and battery in parallel with Z’s capacitor.
The left end of the schematic shows the generator which is connected to the ICE. This end is considered the power flow input. The
term rectifier is also used since the input converter may be used as a passive rectifier. Z-Source network in the center and a CSI at the
output

24

The third circuit architecture is the Input Rectifier/CSI to Output VSI with Z-Source
impedance network and the battery is in parallel with Z’s capacitor. This topology is the
same as directly above accept that the converters are swapped end for end. In this case,
the input converter is the CSI which allows only one directional current flow and bidirectional voltage blocking while the output converter is the VSI converter.
schematic is now shown below.

25

Its

Generator Inverter
Sas-p Sbs-p Scs-p

Rs Ls
Generator /
Starter Motor

+ L1 -

+
VB1
-

+ +
C1C2
- -

as
bs

VG

cs
Sas-n Sbs-n Scs-n

- L2 +

Traction Inverter
S1T-p S2T-p S3T-p
Traction
LT RT Motor / Regenerative
Brake

+
Vi
-

S1T-n

Vm

S2T-n S3T-n

Figure 11 Input Rectifier/CSI to Output VSI with Z-Source impedance
Network and battery in parallel with Z’s capacitor.
The left end of the schematic shows the generator which is connected to the ICE. This end is considered the power flow input. The
term rectifier is also used since the input converter may be used as a passive rectifier. Z-Source network in the center and a VSI at the
output

26

The fourth circuit architecture is the Input Rectifier/VSI to Output CSI with Z-Source
impedance network and the battery is in parallel with the Input Converter. This is like
the second Z-Source circuit shown above except that the battery position is different.
The schematic is shown below.

27

Generator
Inverter
Sas-p Sbs-p

Traction
Inverter
S1T-p S2T-p

Scs-p

S3T-p

+ L1 Generator /
Rs Ls
Starter
Motor

VG

as
bs

+
VB1
-

+
C2
-

+
C1
-

Traction
Motor / ReLT RT generative
Brake

+
Vi

Vm

-

cs
Sas-n Sbs-n Scs-n

- L2 +

S1T-n S2T-n

S3T-n

Figure 12 Input Rectifier/VSI to Output CSI with Z-Source impedance
Network and battery in parallel with Input Converter.
The left end of the schematic shows the generator which is connected to the ICE. This end is considered the power flow input. The
term rectifier is also used since the input converter may be used as a passive rectifier. Z-Source network in the center and a CSI at the
output

28

3. EVALUATION OF MAJOR HEV COMPONENTS AND
OVERALL OPERATION OF HEV

One of the most effective features of the hybrid vehicle is its ability to decouple the
angular velocity of the ICE from the velocity of the vehicle. The extent of which a
particular hybrid is able to do this allows the ICE operating point to be kept within its
most efficient operating points. . It is the hybrid equipment (generator, inverters, battery
and converter, and traction motors) that allow the decoupling of ICE torque and speed
from that of the traction wheels. This equipment also allows the recapture of braking
energy which would be lost with non-hybrid vehicles.

29

ICE

Generator

Battery
Power
Electronics

Motor /
Generator

Wheel

Gear Box

Wheel

Figure 13 Series Hybrid Electric configuration
The traditional vehicle has a transmission that allows a decoupling of ICE torque speed
from vehicle speed to some degree. Due to the transmissions several discrete gear ratios,
the ICE can operate at one torque and speed point and the vehicle wheels can operate at a
different torque-speed point; however, the torque speed decoupling is not complete with
the traditional transmission. A specific torque speed point can also be referred to as a
specific power point. In general, the power point of the ICE is equal to the power point
of the vehicle wheels (neglecting losses). A particular power point can be achieved with
an infinite number of torque-speed combinations most of which occur at low efficiency
and just a small grouping of such points is at high efficiency. It is then desirable to
completely decouple the ICE torque-speed operating point from that of the vehicle
30

wheels so that the ICE can remain operating in its most efficient region. The hybrids are
able to achieve this decoupling to a greater extent that the traditional transmission and
each different hybrid type (mild hybrid, full hybrid) is able to improve this decoupling to
lesser or greater degrees. The series and series/parallel full hybrids are capable of
completely decoupling the torque-speed operating point of the ICE from that of the
vehicle wheels and the ICE power and vehicle wheel power is also decoupled to the
extent that the battery can source or sink the power difference. If power from the ICE is
different than power to or from the traction motors, this difference of power can be
provided by or absorbed by the battery. Also, it is the battery that also allows braking
energy to be stored. The main components that will be discussed in this section are the
ICE, generator, battery, inverters (controller in the figure above), and traction motors.
Each of these components has its own set of characteristics and it is desirable to operate
then in an efficient manner. In the following sections, each of these major components is
discussed and analyzed to determine what constraints they impose on the overall design
of the hybrid topology and how these constraints can be managed in a more efficient
manner. First, we consider the ICE of the hybrid vehicle.

3.1 Characteristics of the Internal Combustion Engine (ICE)

Below is a torque vs. angular velocity graph of a typical diesel ICE. Laid on top of this
graph are the efficiency maps of the ICE. The inner most loop (from 1212 to 1575
revolutions per minute and 900 to 1650 Newton Meters of Torque) is the operating region
of highest efficiency. Operating points outside of this region are of lesser efficiency and
31

one can see that efficiency begins to decrease rather quickly for operating points further
outside of this region.

1800
1600

Torque[Nm]

1400
1200
1000
800
600
400
200
800 1000 1200 1400 1600 1800 2000 2200 2400

Ne[r*min-1]
Figure 14 Typical Torque Vs. Speed curves of a diesel ICE.
Highest efficiency operation is shown in the center most enclosed line spanning the
region from 1212 to 1575 revolutions per minute and 900 to 1650 Newton Meters of
Torque.

For a diesel engine, this region may achieve efficiency as high as 42% with a target
efficiency of 50% by the year 2015[28]. Since the efficiency of the ICE is sensitive to its
torque – speed operation point, a significant amount of savings can be achieved by
ensuring that the ICE operating point does not deviate from this region. Since the
generator is connected to the ICE, it should be designed to operate at the ICE’s optimum
32

speed and to produce the desired voltage at that speed. Next the hybrid battery is
discussed.

3.2 Characteristics of the Hybrid Battery Pack

The hybrid vehicle uses several power electronics components. Such components are an
AC to DC inverter connected to the generator, an AC to DC inverter connected to the
traction motors and a DC/DC converter connected to the battery and to the inverters. See
the following schematic.

Generator Inverter

DC-DC
Converter

Traction Inverter

G

M

C1
Vd
LVBatt

Figure 15 Traditional Back to Back VSI / VSI hybrid topology

The AC to DC inverters and DC/DC converter are connected together through a common
set of conductors called the DC-Link. It is typical that the battery voltage is different
from that of the DC-Link for several reasons. As such, the battery requires a DC/DC
converter to control the difference in voltage levels between that of the battery and of the
33

DC-Link. Below is a discussion about the batteries and issues related to adding more
battery cells in series and in parallel.

3.2.1 Comparison of the high voltage battery vs. the low voltage battery.
High voltage battery option.
The high voltage battery has the advantage of being able to connect directly to the DCLink without the use of a DC-DC converter. This assumes that the DC-Link is held at a
constant voltage level and not all hybrids operate in that manner. The high voltage
battery option is simple in design and uses fewer electronic components.

It does,

however, require a greater number of battery cells to be stacked in series. There is also a
fundamental problem with connecting batteries in series. If the Lithium Ion battery is
used as an example, due to its lighter weight and relatively fast charging rate, e.g.
Valence Technologies U27-FN130, it can receive a maximum charge current of 100A. A
series connection of 55 batteries is needed to achieve a total voltage of 700 volts. With
such a series connection, a failure of any one of the 55 batteries will cause a serious
degradation of performance.

The reliability equation for series-parallel arrays of

electronic components is as follows.

R(t )  [1  p n F (t ) n ]m  [1  (1  qF (t ) n )]m

Where:


R(t) is the reliability,
34

(eq 3.1)



p is the conditional probability of a short,



q is the conditional probability of an open,



m is the number of parallel paths,



n is the number of components in series,



F(t) is the cumulative distribution function of time to
failure of series components [19]

In this particular case, the short circuit failure rate is quite low for batteries. Typically,
battery failures occur from an increase in resistance and or a decrease in voltage. As a
result, reliability decrease is much more likely to occur from series connections than from
parallel connections.

If it is necessary to receive a charge at greater than 100 A, additional parallel
combinations of 55 series batteries will be needed. As can be seen, the number of
batteries required can add up quickly. Also, with series connections, each battery is
slightly different than the other batteries that it is connected to (different internal
resistance, different voltage level, etc.) As a result, some of the series batteries are
putting out more power than others. Under high current levels, a poorly functioning
battery may actually be dissipating more power than it is providing. See Figure 16. The
net result is a compromised performance of the whole stack.

35

Battery 1

Vs_1

+
_

Ri_1

+
+
_

Battery 2

Vs_2

Vbatt_1

_
R_load

+
+
_

Ri_2

+
_

I_load

Vbatt_2

_

Figure 16 Series Combination of Battery Cells
If I_load is large and Vs_1 is small, Vbatt_1 can be negative and
Battery 1 can be absorbing power from Battery 2.
Where Ri is the internal resistance of a battery, Vs is the internal voltage
potential of a battery, Vbatt represents one battery unit

As the age of the batteries increase, these conditions worsen. Due to these factors, the
high voltage battery has a lower reliability than the DC-DC boost converter and low
voltage battery combination. Battery management electronics can be added to monitor
individual battery condition and allow for by-passing weak cells, but this adds to the
complication and cost.

Typically, a high voltage series stack is more costly than a low voltage stack of the same
Amp hour rating. Also with the high voltage battery stack comes the additional safety
concerns for service personnel, rescue workers, etc.

36

3.2.2 Comparison of the high voltage battery vs. the low voltage
battery. Low voltage battery with DC/DC converter option.
With a low voltage stack, a DC-DC boost converter is required. Since the low voltage
stack voltage level is less than that of the DC Link, a boost converter is required between
the DC Link and the batteries. The disadvantage of this low voltage arrangement is that
this additional circuit (DC-DC boost converter) is needed. Along with the additional
circuitry come additional cost, complexity, and lowering of reliability, due to the
additional components. The advantages of this design however, is that the total number
of batteries in series will decrease, and this will increase system reliability for the reasons
discussed above.

The use of a DC-DC converter adds an extra level of control for the battery SOC by
controlling the DC output voltage level. Although this level of control is not needed to
control battery SOC since coordinated control of the generator inverter/rectifier and
traction motor inverter/rectifier also achieve SOC control. Still, the DC-DC converter
can provide this as well as an added level of battery health monitoring and control. The
low voltage option also reduces voltage levels present when the vehicle is not running
which can reduce risk factors to passengers and emergency personnel in the case of a
collision involving the vehicle.

37

3.2.3 Summary of High Voltage Battery Option vs. Low Voltage
Battery Option
High Voltage Option
Pros:
 Battery Voltage = DC Link Voltage and a DC-DC converter is not
Required
Cons:
 Greater number of Batteries in series and a related decrease in reliability
of the battery stack
Low Voltage Option
Pros:
 Lower number of batteries in series and a related increase in reliability of
the battery stack.
 Extra level of control for battery SOC
 Reduced non-operating voltage levels
 Reduced energy loss during low power operation
 Reduced THD during low power operation
 Reduced total number of cells
Cons:
 A DC-DC converter is required
 Decreased reliability due to additional components.

For these reasons, a lower voltage battery stack is usually preferred.

3.3 Characteristics of Inverters and Converters

Perhaps the most significant component of the inverters and converters is the switching
device. For inverters and converters of this power rating, this device is the IGBT module.

38

Figure 17 IGBT module used in inverters and converters

The IGBT module consists of an insulated gate bipolar junction transistor and a PN
junction diode. The losses of these devices are of concern to this power loss analysis.
For example, the following equations show the switching losses and conduction losses of
an inverter’s IGBT and diode. In these equations, notice that each of the switching losses
(Pon, Poff, and Prr) are proportional to Vcc which is the DC-Link voltage. The IGBT
conduction losses Pi and diode conduction losses Pd are not related to the DC-Link
voltage.

39

I cM 2
1
Pon  Vcc  trN
Fs
8
I cN

(eq 3.2)

 1
Poff  Vcc  I cM 2  t fN  Fs 
 8 I
cN








 0.1404  0.1902 I cM  0.0075 I cM

I



I cN
 cN
Prr  FsVcc 
 
I
0.8 
  0.05 cM 
 I cM  trrN
I cN
 
 

 

(eq 3.3)







Q 
 rrn 







2

(eq 3.4)

1 M
 V  Vco
 1 M

Pi   
cos( )  ceN
I cM 2  
 cos( ) Vco I cM
 8 3
 I cN
 2 8

(eq 3.5)

1 M
 V  VFo
 1 M

Pd   
cos( )  FN
I cM 2  
 cos( ) VFo I cM
I cN
 8 3

 2 8

(eq 3.6)
[25]
Where:
Pd = diode conduction loss
Pi = IGBT conduction loss
Poff = Turn off loss of the IGBT
Pon = Turn on loss of the IGBT
Prr = reverse recovery loss of the diode
Fs = switching frequency
IcM = maximum collector current
40

IcN = rated collector current
M = modulation index
Qrrn = rated recovery charge
tfN = rated fall time

trN = rated rise time
trrN = rated recovery time
Vcc = DC bus voltage
VceN = rated collector to emitter voltage
Vco = threshold collector emitter voltage
VFN = rated diode forward voltage
VFo = threshold diode forward voltage

The significance of these equations will become more apparent later in this discussion.
Next, the traction motors will be discussed.

3.4 Traction Motor Operating Conditions

Below in Figure 18, are images that represent the operating speed of city (urban) driving.
These figures are prepared by the United States Environmental Protection Agency (EPA)
and are based upon their studies of typical vehicle operating conditions. This EPA chart
shows the Urban Dynamometer Drive Schedule.

41

Figure 18 EPA Urban Dynamometer Driving Schedule (UDDS)

This chart shows that the operating velocities of urban driving are far below full speed,
for the vast majority of time. Since the city or commuter bus operates with driving
patterns similar to those of the EPA UDDS, this chart is relevant to this analysis.

The traction motors are coupled to the traction wheels through a constant ratio gear box
and therefore, the traction motor angular velocity is directly related to the vehicle
velocity. Therefore, the traction motor angular velocity required is much below full

42

speed for the vast majority of time and the line to line voltage required by the motor to
achieve the desired current is much lower than full voltage for the vast majority of time.

3.5 Summary of Hybrid Major Components

The major components of the hybrid vehicle have been reviewed and the relevant
operating conditions and characteristics of each component have been considered.
Considering the traditional hybrid topology, shown here again for convenience, the
component characteristics will be summarized from left to right.

Generator Inverter

DC-DC
Converter

Traction Inverter

G

M

C1
Vd
LVBatt

Figure 19 Traditional Back to Back VSI / VSI hybrid topology
Starting with the ICE, not shown above, it was determined that a constant operating speed
would be the most efficient for the ICE. As a result of this, the generator will then be
operated at a constant speed and therefore a constant voltage. In general, generators can
be designed to be more efficient if they generate at high voltage and low current. The
greatest gains in efficiency can be achieved by catering to the ICE’s efficiency sensitivity
43

to torque and speed operating points. Therefore, the generator will output a constant high
voltage. This in turn sets the DC-Link voltage to be high.

The battery voltage desired is significantly below the voltage level of the DC-Link as
previously discussed, and the voltage output of the traction inverters will much below full
voltage for the vast majority of vehicle operating points since the vehicle operating
speeds were shown to be well below full speed for this amount of time.

Traction motor line to line voltage can be lowered by lowering the modulation index (or
decreasing vector magnitude when using vector control); however, we also saw that the
switching losses of the inverter switching devices (IGBT and freewheeling diode) were
directly proportional to the DC-Link voltage. With this in mind, it is desirable to keep
the voltage across the switching devices of the traction inverter (TI) and DC/DC
converter as low as possible. This could be done by outputting a low voltage from the
generator and then increasing the generator voltage as need by the traction motor.
Raising this voltage can be achieved in two ways. The first if to initially operate the ICE
at a low speed and then increase this speed as necessary, but as discussed previously, this
in not desirable based upon the ICE’s efficiency map. Another method is to generate at
low voltage and constant speed and then boost the voltage by lowering the GI modulation
index. This method would cause generating to be done at low voltage for the majority of
time and therefore at higher current than necessary which is generally not efficient. Also,

44

this uses the generator’s inductance to perform the boost function and this also raises the
losses encountered by the generator.

In summary, it is desired that the generation be held at a constant high voltage, that the
battery voltage remain low, and that the voltage across the TI be variable and controlled
to be as low as possible for the present velocity of the bus. This is not possible with the
traditional topology. The DC-Link of the traditional topology (see Figure 19) is common
to both the GI and TI and thereby imposes that the voltage across the GI is at all times
equal to the voltage across the TI. Hence, a new topology is needed to achieve the more
favorable operational conditions.

45

4. PRESENTATION AND BRIEF ANALYSIS OF
PROPOSED TOPOLOGIES
Being able to operate the generator inverter’s DC-Link voltage at a high voltage and the
traction inverter’s DC-Link voltage at a low voltage provides very significant gains in
efficiency. Further, it should be pointed out that that the traction inverter must handle the
total vehicle power (generator and battery power) which causes an increase in efficiencies
of the traction inverter to be even more significant.

What is desired is a topology that can allow the ICE to operate in its most efficient torque
and speed region, a generator designed to operate at high voltage and low current at this
Traditional Back to Back VSI /
speed, and a DC-Link that can offer as low of a DC voltage to the traction invert as is
dictated by constant volts/Hz operation of the traction motors.

Ideally, the new topologies with lower TI losses, would not incur any additional losses in
the process of decoupling the GI voltage from the TI voltage. However, the process of
providing lower voltage to the TI does incur additional conversion losses. The trick is to
incur less additional conversion losses than is saved by the reduced losses in the TI. That
is, even though the process of decoupling the GI DC-Link voltage from the TI DC-Link
voltage causes some additional conversion losses, the goal is to achieve a greater
reduction of losses in the TI which will more than compensate for this and yield a net
reduction in losses.

The following sections 4.1 through 4.8 describe the proposed new topologies. Chapter 5
46

will show the power losses of these new topologies as compared to the traditional
topology.

4.1 VSI / VSI with Conventional Mid-Link DC / DC Converter

G

M

Figure 20 VSI / VSI with Conventional Mid-Link DC / DC Converter

The topology shown in Figure 20 has as its basic structure a generator VSI and traction
motor VSI. The Low voltage battery is coupled to the DC-Link through two DC/DC
converters which boost from the battery toward the GI and toward the TI and buck from
the GI toward the battery and from the TI toward the battery. This topology offers a
Generator inverter DC voltage that is independent from the DC voltage at the traction
inverter. The TI DC-Link voltage can be as high as VBATT

1
and as low as VBATT.
1 D

Only the generator power flows through the generator inverter and through the left
DC/DC converter whereas both the generator power and the battery power flows through
the right DC/DC converter and the TI. With this topology, the generator can be designed
for high voltage and low current. The battery can be a low voltage battery stack due to
the GI’s DC/DC converter. The traction inverter’s DC-Link voltage (VDC_TI) can range

47

from the battery voltage (VBATT) to the VDC_TI required for maximum vehicle speed.
VDC_GI and VDC_TI are independent of each other and only dependent on VBATT. The
topology allows the more efficient high voltage low current generator design and the
more efficient low VDC_TI across the switching devices of the TI for operation at less
than maximum vehicle speed. The following diagram summarizes the per unit power that
must flow through each section. Where possible, the goal is to implement the least
efficient converters where the power flow is lowest and to implement the most efficient
conversion processes where power flow is greatest. In this case, however, the DC/DC
conversion on the GI side (with ½ per unit power flow) is just as efficient as that on the
TI side (with 1.0 per unit power flow).

SPU = 1/2

SPU = 1/2

SPU = 1

G

SPU = 1
M

Figure 21 VSI / VSI with Conventional Mid-Link DC / DC Converter.
Per unit power flow through each inversion or conversion process is indicated.
State of charge (SOC) of the batteries can be controlled by controlling the relative power
flow from the generator, to or from the traction motors, and or controlling the switching
duty cycles of the DC / DC converters.
48

4.2 VSI / VSI with “H” Bridge Coupled Battery Pack
The topology shown in Figure 22 is based upon the well-known VSI /VSI configuration
where a VSI is used at both the generator end and traction motor end of the topology. In
this particular case, the battery pack can be connected to the DC-Link in two ways, or not
at all.

C

G

S1

S2

M
S1'
S2'
VBatt
VDC_G

VF

I

Figure 22 VSI to VSI topology with an “H” bridge coupling the battery
pack to the DC-Link

If switches S1 and S1’ are gated on simultaneously, the voltage across the LC filter at the
TI is:
VF = VDC_GI - VBATT

(eq 4.1)

If S2 and S2’ are gated on simultaneously:
VF = VDC_GI + VBATT

(eq 4.2)

If S1 and S2’ are gated on simultaneously or S1’ and S2:
VF = VDC_GI

( eq 4.3)

49

Because of the inductor of the LC filter at the TI end, the voltage across the TI can be
controlled to be any voltage level between (eq 4.1) and (eq 4.2).

The following figure shows the per unit power that must be managed by each inverter
and converter.

SPU = 1/2
G

SPU = 1/2
S2

SPU = 1

S1
M

S1'

S2'

Figure 23 VSI to VSI topology with an “H” bridge coupling the battery pack to the
DC-Link.
Per unit power flow through each inversion or conversion process is indicated.
Notice that while S1 and S1’ are gated on, that the battery is being discharged. While S2
and S2’ are gated on the battery is being charged and in the third and fourth states, the
batteries are neither charged nor discharged. As a result, the state of charge (SOC)) of
the battery pack can be controlled. However, the SOC of the battery and the voltage
across the TI cannot be effectively controlled at the same time which invalidates this
topology.

A second issue with this new topology is that regenerative braking will not work at low
speeds when VDC_TI < VBATT.
50

4.3 VSI / VSI Structure, LV Battery with Boost Converter, & DC-Link
with Buck Converter

G

(+)

M

(-)

Figure 24 VSI / VSI Structure, LV Battery with Boost Converter, & DC-Link with
Buck Converter
The topology shown in Figure 24 uses the standard VSI / VSI platform. Added to this
topology are a boost converter for the battery pack and a buck converter between the
battery converter and the TI. This configuration allows the DC-Link voltage at the GI to
be held constant and allows the DC-Link voltage at the TI to be varied from zero to the
voltage level required for maximum traction motor speed.

VDC _ TI  VDC _ GI  D

(eq 4.4)

Hence the traction motors can be operated via constant volts/Hz control while holding
the TI modulation index at 1.0. Holding the voltage across the TI to the lowest level
needed to support the required output voltage to the motors greatly reduces the TI
switching losses. Holding the modulation index of the TI at 1.0 reduces the harmonic
content of the inverter output and thereby reduces losses in the traction motors as well.

51

M

(+)

G

(-)
SPU = 1/2

SPU = 1/2

SPU = 1

SPU = 1

Figure 25 VSI / VSI Structure, LV Battery with Boost Converter, & DC-Link with
Buck Converter.
Per unit power flow through each inversion or conversion process is indicated.
The kVA rating of the battery’s boost converter only needs to be rated for the battery
power; however, the kVA rating of the TI’s buck converter must be rated for total system
power.

4.4 VSI / VSI Structure, LV Battery with Buck-Boost Converter, & DCLink with Buck Converter

S3
S4

(-)

M

VDC_TI

(+)

(+)

VDC_GI

G

(-)

Figure 26 VSI / VSI Structure, LV Battery with Buck-Boost Converter, & DC-Link
with Buck Converter

52

The configuration shown in Figure 26Figure 26 uses the traditional back to back VSI /
VSI topology. Added to this topology is a buck converter on the GI side that bucks
voltage from the GI toward the TI and boosts voltage from the TI toward the GI. This
converter can buck from the GI voltage VDC_GI all of the way to zero volts. The low
voltage battery is connected to the DC Link on the TI side via a buck-boost converter that
is capable of bucking voltage all the way down to zero volts and boosting voltage up to
infinity (theoretically).

 D 
VDC _ TI  VBATT 
 (1  D) 




(eq 4.5).

This configuration allows the complete decoupling of voltage levels from the GI end of
the DC-Link to the TI end. The DC-Link voltage level at the TI can be varied from 0V to
as high as is needed for constant voltz/Hertz operation, while at the same time, the DCLink voltage at the GI can be held constant.

(+)

Spu = 1/2

(+)

(-)

G

(-)

Spu = 1/2

Spu = 1/2

M

Spu = 1

Figure 27 VSI / VSI Structure, LV Battery with Buck-Boost Converter, & DC-Link
with Buck Converter.
Per unit power flow through each inversion or conversion process is indicated.
53

Regarding power flow, only the generator power must flow through the first converter on
the GI end and only the battery power must flow through the second (battery) converter.
All system power must flow through the TI. Also with this topology, the buck-boost
converter (connected to the battery) switching devices must be oversized as compared to
a boost or a buck converter. The voltage seen across S3, and S4 will be
VS3 = VS4=VDC_GI + VBatt

(eq 4.6)

The buck-boost converter allows the complete de-coupling of VDC_GI from VDC_TI, but
this large voltage across S3 and S4 also causes significant switching losses.

4.5 VSI / VSI Structure, LV Battery, DC-Link with Integrated Magnetic
Boost Converter Using Open Circuit Zero States of Inverter Phase Legs

G

VDC_TI

VDC_GI

M

VBatt
Figure 28 VSI / VSI Structure, LV Battery, DC-Link with Integrated Magnetic
Boost Converter
Using Open Circuit Zero States of Inverter Phase Legs.
The topology shown in Figure 28 is also based upon the basic VSI / VSI configuration.
The connection from the DC-Link voltage on the GI end is made via an integrated
magnetic DC-DC converter. Likewise, the connection from to the DC-Link voltage on
54

the TI end is made in the same way. Theoretically, the GI DC-Link voltage could be
controlled by controlling the open circuit zero states (OCZ) of the GI. Since the GI has to
use both active states and OCZ states in its normal function to control the generator, the
DC/DC conversion between the GI and the battery pack can use these zero states. The
zero states are traditionally evenly divided between closing all three upper switches, for
the open circuit zero state (upper zero state), or all three lower switches (lower zero
state). By choosing which zero state to use, the DC-Link voltage could be controlled.
When the upper switching devices are used to perform the zero state, the inductors see
the following voltage.

VL  VDC _ GI  VBatt

(eq 4.7)

And when the lower switches are used for the zero state, the inductor voltage is:

VL  VBatt

(eq 4.8)

As can be deduced, the power flow to and from the GI and battery can be controlled by
controlling the duty cycles of the upper or lower zero state.

Voltage conversion from the battery pack to the DC-Link at the TI end uses the same
method. Theoretically, the DC voltage at the TI can range as low as the voltage of the
battery pack to infinity. This topology benefits from reduced components by not having
a separate DC/DC converter.

55

G

M

Spu = 1/2

Spu = 1

Figure 29 VSI / VSI Structure, LV Battery, DC-Link with Integrated Magnetic
Boost Converter
Using Open Circuit Zero States of Inverter Phase Legs. Per unit power flow through
each inversion or conversion process is indicated.
Only the power of the GI must flow through the left converter whereas both the battery
power and the GI power must flow through the second converter. Below, in Figure 30,
are some simulation results under the condition that only the upper open circuit zero state
is being used. The first image shows switching states (phase “A” upper switch, phase
“A” lower switch, … ,phase “C” lower switch). The second image shows the back emf
from the traction motor and the third image is the DC-Link voltage being held at 100V
(the simulation battery voltage).

Figure 31 shows the same switching waveforms and DC-Link voltage with a zoomed in
time frame.

56

A upper
A lower
B upper
B lower

Switch
State

C upper
C lower

200.0
(V)

Vdc

100.0
0.0

(V)

EMF a to b

150.0
100.0
50.0
0.0

EMF b to c

EMF c to a

-50.0
-100.0
-150.0
57.58m

57.60m

57.62m

57.64m
t (s)

57.66m

57.68m

57.70m

Figure 30 Inverter switching state, DC-Link voltage, and back EMF
of the VSI / VSI Integrated Magnetic topology
using only upper OCZ states. Vbatt = 100V in this simulation.

57

A upper
A lower
B upper
B lower
C upper
C lower
200.0

(V) 100.0

0.0
57.58m

57.60m

57.62m

57.64m
t (s)

57.66m

57.68m

57.70m

Figure 31. Switch states and DC-Link voltage.
Same conditions as Figure 30. Zoomed in on the time scale. High switch state =
switch closed, low switch state = switch open.
This topology has some problems however. Notice that the line to line voltage of the
motor and generator are across the boost inductors at all times. As such, the topology is
not practical for low motor speeds. Therefore, this topology will be excluded from
further study.

58

4.6 VSI / VSI Structure, LV Battery, DC-Link with Bi-Directional Buck
and Boost Converter

G

M

Figure 32 VSI / VSI Structure, LV Battery, DC-Link with Bi-Directional Buck
and Boost Converter

This configuration also is based on the traditional VSI / VSI structure, but with two buckboost converters on the DC-Link. The buck boost converters are of an “H” bridge type.
Each converter has four switches rather than the two switches of other designs; however,
this buck-boost converter does not invert the polarity and the voltage stress on the
switches is lower. With these two buck-boost converters, the voltage level at the TI end
and GI end of the DC-Link are completely decoupled. The voltage level from the GI can
be varied from zero to infinity (theoretically) and the same is true for the TI end of the
DC-Link.

59

G

M

Spu = 1/2

Spu = 1/2 Spu = 1

Spu = 1

Figure 33 VSI / VSI Structure, LV Battery, DC-Link with Bi-Directional Buck
and Boost Converter.
Per unit power flow through each inversion or conversion process is indicated.
The left converter must conduct just the power flow of the generator while the second
converter (on the right) must conduct both the generator power and the battery power
(total system power). This topology has the advantage of being able to completely decouple VDC_GI from VDC_TI, but with the disadvantage of even more additional
components.

60

4.7 Back to Back Current Fed Quasi-Z-Source Inverters, LV Battery
S2
D1
S1

D2

VDC_TI

M

VBatt

VDC_GI

G

= RB-IGBT

Figure 34 Back to Back Current Fed Quasi-Z-Source Inverters, LV Battery

Figure 34 shows the back to back current fed quasi “Z” source inverter topology. Notice
that the GI and TI are using a reverse blocking IGBT. This device can block voltage in
both directions but will allow current to flow in only one direction when gated on. It also
uses the quasi-Z-source impedance network composed of two inductors and two
capacitors. The quasi-Z-source impedance network on the generator end includes an
IGBT module to allow power flow from the battery to the generator/starter motor to start
the diesel engine. The impedance network on the TI end includes just a diode. More
about the operation of this topology will be covered in the last chapter.

G

M

Spu = 1/2

Spu = 1

= RB-IGBT

Figure 35 Back to Back Current Fed Quasi-Z-Source Inverters, LV Battery.
Per unit power flow through each inversion or conversion process is indicated.
61

The impedance network on the GI end of the topology only carries ½ of the total system
power and the corresponding impedance network on the TI end carries full system power.
Also, this topology benefits from not having a separate DC/DC converter for the battery,
but does have additional inductors, switching devices and switching states.

4.8 Back to Back Voltage Fed Quasi-Z-Source Inverters, LV Battery

G

M

Figure 36 Back to Back Voltage Fed Quasi-Z-Source Inverters, LV Battery

This configuration is based upon the VSI / VSI topology with two voltage fed continuous
current qZSIs added to the DC-Link. The left (generator) qZSI bucks the higher voltage
at the generator end of the DC-Link down to the battery voltage, and boosts the lower
battery voltage up to the voltage level at the generator end of the DC-Link. Likewise, the
right (traction) qZSI can boost the voltage level at the traction inverter end of the DCLink from the battery voltage up to the level needed for full speed operation of the bus.
The voltage level at the generator end can be as high as is desired. The GI short circuit
zero states buck the generator voltage down toward the battery. The DC-Link voltage at
62

the TI end can be controlled to range from the battery voltage to infinity (theoretically).
The voltage level at the TI end of the DC-Link can be boosted via the short circuit zero
states of the TI. The continuous current voltage fed qZSI can allow bi-directional power
flow by using bi-directional current, unidirectional blocking switches such as shown in
Figure 36. To flow power from the generator to the battery, the IGBT needs to be gated
on. To flow power from the battery to the TI, the IGBT is gated open.

G

M

Spu = 1/2 Spu = 1/2 Spu = 1

Spu = 1

Figure 37 Back to Back Voltage Fed Quasi-Z-Source Inverters, LV Battery.
Per unit power flow through each inversion or conversion process is indicated.
Just the power flow from the GI flows through the left qZSI network, and both the battery
power and the generator power flow through the right qZSI network. This topology also
benefits from not needing a separate DC/DC converter for the battery, but has additional
inductors, switching devices and switching states.

63

4.9 Summary of proposed topologies
The topologies shown in section 4.2 VSI / VSI with “H” Bridge Coupled Battery Pack
and section 4.5 VSI / VSI Structure, LV Battery, DC-Link with Integrated Magnetic Boost
Converter Using Open Circuit Zero States of Inverter Phase Legs have been shown to be
dysfunctional in their present form. They will be excluded from the analysis in the
following section. The various topologies presented are able to decouple the voltages to
lesser or greater degrees. Some are able to completely decouple the voltage and others
are able to achieve a variable voltage at the TI that varies from the battery voltage up to
the max. Each topology has its own pros and cons regarding the distribution of power
flow through components, voltage stresses seen by switching devices during the open
state and switching transition, additional inductors, presence or absence of battery
DC/DC converter, and additional switching states. These factors will all be analyzed in
the following section.

64

5. LOSS ANALYSIS OF SELECTED TOPOLOGIES

This section presents analytical results of the new topologies.

Of the topologies that

were introduced in chapter 4, those which are functional are considered here in more
detail. The losses of switching devices in the inverters and converters as well as the
inductor losses are calculated in the following sections. Capacitor losses are ignored in
this study because the capacitor losses of the traditional topology (which provides the
base line or point of comparison) are essentially equivalent to the capacitor losses in the
new topologies and therefore can be neglected in this comparative study. The next
section describes the formation of the comparative base line.

5.1 Base Line:
First, the efficiency of the traditional hybrid bus topology will be analyzed to form a base
line against which other topologies proposed in this paper can be compared. Turn on
loss, Pon, turn off loss, Poff, IGBT conduction loss, Pi, Diode conduction loss, Pd, and
diode reverse recovery loss, Prr, of the inverters and converter were calculated in a
manner similar to that in [25 ].

5.1.1 Evaluation of Efficiency of the topologies.
The equations used to calculate the efficiencies of the topologies were briefly referred to
before, and are discussed more thoroughly here. Figure 38Error! Reference source not
found. below shows the voltage and current of a typical semiconductor switching device
during the turn on transient, conduction, and turn off transient. Notice near the left end of
the picture, that while the collector current (IC) of the IGBT is increasing, that the
65

voltage across the IGBT remains high. This is the turn on transient of the device. It is
the presence of the current through the device and the simultaneous voltage across the
device that causes a power loss. The time period, tr, is the rise time for the current which
results in the turn-on switching losses. Since the voltages and currents are changing, the
product of these values must be integrated during this time period to calculate the energy
lost. Also, we desire to cumulate the switching losses during a fundamental power cycle
and average them to find the rate of energy loss known as power loss. This must be done
for the turn-off transient during tf, and for the conduction period between tr and tf. In
general, the fundamental power cycle of the inverters can be found by following the
sinusoidal current. The DC/DC converter requires different equations since their power
flow does not follow a sinusoidal current waveform. Also, the Z-source type topologies
have additional switching states (short circuit zero or shoot-through) that must be
considered.

66

Voltage (V), & Current (I)

VCE

VCE
Irr

IC

trr
tr

IC
VCE

IC

Time, (t)

tf

Figure 38. Voltage and current waveforms of a typical semiconductor switching
device
The loss comparison between the new topologies and the traditional topology were
calculated using the following equations as taken from [25].

I cM 2
1
Pon  Vcc  t rN
Fs
8
I cN

1 M
 V  Vco
Pi   
cos( )  ceN
I cM 2 
I cN
 8 3

 1 M

 cos( ) Vco I cM

 2 8


(5.1)

(5.2)

1 M
 V  VFo
Pd   
cos( )  FN
I cM 2 
I cN
 8 3

M
 1


cos( ) VFo I cM

8
 2


(5.3)

A slight modification is made to the equations found in [25] for Poff and Prr as was
67

necessary for this particular case. The equations used for Poff and Prr are presented just
below.

 1
Poff  Vcc  I cM 2  t fN  Fs 
 8 I
cN







(5.4)



I
0.8 
 I cM  t rrN 
Qrrn   0.05 cM 

I cN
 





(5.5)

where:

Pon is the IGBT turn on loss,
Poff is the IGBT turn off loss including tail current loss,
Pi is the IGBT conduction loss ,
Pd is the freewheeling diode conduction loss,
Prr is the diode reverse recovery loss,
Vcc is the dc-link voltage,
trN is the rated rise time,
IcM is the maximum collector current,
IcN is the rated collector current,
Fs is the switching frequency,
M is the modulation index,
θ is the power factor angle,
VceN is the rated voltage between the collector and emitter,
Vco is the threshold voltage between the collector and
emitter,

VFN is the rated diode forward voltage,
68

VFo is the threshold diode voltage,
tfN is the rated fall time,
Qr r n is the rated recovery charge.

These variables can be taken directly from IGBT module specification sheets. The
equations used to calculate the switching losses of the traditional configuration’s dc/dc
converter are:

Pon 

t
1
 Vcc I dc ) 2 rN Fsdc
2
I cN

(5.6)

 V

 Vc 0
Pi _ DC   ceN
I dc  Vc 0  I dc


I cN





t fN
1
Poff  Vcc I dc 2 
 Fsdc
2
I cN
Ptail _ DC 

(5.8)

3
Vcc I dct tail Fsdc
80

[((

)

(5.7)

(5.9)

)

]

(5.10)



i
0.3 

Prr _ DC  Vcc t rrN  I dc 2  0.045 rrN2 

I cN 

I cN







I
I dc  0.225 rrN  0.8   0.281I rrN  Fsdc


I cN





69

(5.11)

Where Idc is the DC current flowing through that device, and this depends upon which
side of the converter the devices is on (high voltage or low voltage side), and Fdc is the
switching frequency of the DC/DC converter.

Inductor loss was approximated by the following equation 5.13 which has provided a
reasonable loss approximation for inductors tested in the lab.

Pind _ dc / dc  Pdc  0.0043( Bdc / dc  1)  0.0057

(5.13)

Where:

Pind_dc/dc is the power loss in the DC/DC converter inductor
Pdc is the DC power being absorbed or provided by the battery
pack

Bdc/dc is the boost factor of the DC/DC converter.

For inductors sized and designed for the relevant power levels of this study, this equation
has provided more accurate and slightly higher loss values than those of the Steinmetz
equation. Therefore, by using equation 5.13, we err on the side of caution. This means
that in the comparisons that follow, actual results of the new topologies will be slightly
better than indicated by our conservative graphs.

70

For the comparison, a common set of conditions must be followed for both the new and
traditional topology. The conditions for the comparison are as follows.
Conditions:


Generator line-to-line voltage is 429 Vrms;



rectified generator voltage is 700 Vdc;



TM line-to-line voltage sweeps from 29 Vrms to 429 Vrms;



switching frequency is 10 kHz;



power factor is 90%;



battery pack voltage is 350 Vdc;



switching, conduction and reverse recovery losses are based upon a typical
1200-V/ 450-A IGBT, and diode, or a reverse blocking IGBT (RB-IGBT)
as applicable.



For the traditional topology in figure 1, the TI voltage is at all times held
equal to the GI voltage of 700 Vdc while the TI voltage of the proposed
topologies vary to the extent disclosed in the specific topology’s section
below.

5.1.2 Experimental Results
Before a base line is calculated, some experimental results are desired to confirm the
concept and to verify the validity and accuracy of the loss equations. To achieve these
71

results, a DC-Link, a traction inverter and a traction motor were required. Toward these
ends, a rectifier and inverter were built each using a 6MBI450U4-120 IGBT module six
pack which is rated for 1200V and 450A. The rectifier used just the diodes of the six
pack as a passive rectifier. The inverter also included a custom built gate drive and gate
drive power supply, and both the inverter and rectifier has a 510 micro Farad capacitor
bank. The inverter also has HTB 200-P current sensors on the phase “A” and “B”
conductors and a LM25-P voltage sensor on the DC-Link. The motor being used as the
traction motor is an interior permanent magnet AC motor (IPMAC) which was connected
to a chassis dynamometer. In this case, the vehicle on the chassis dyno. is the load and
the IPMAC motor connected to the chassis dyno. will function as our traction motor. A
TMS320F28035 DSP board loaded with an adapted field oriented control program was
used to control the motor. The digital signal processor (DSP) board was connected to the
inverter via an optocouple board, and signal conditioning circuit boards for the current
sensors, voltage sensor, and rotor position sensor. The experimental set up schematic and
pictures are shown below.

VDC_TI

(+)
Variac
Grid

(-)

Open

Figure 39. Experimental set up schematic diagram

72

Traction
Motor

(a) Rectifier on the left and inverter on the right.

(b) Variac
Figure 40 Experimental Setup

73

One of the adaptations required for the field oriented control algorithm is the addition of
DSP board inputs for the rotor position signal. The rotor position sensor is made up of
three coils wound on part of the stator. One coil is the excitation coil and two coils are the
sensor coils that are coupled to the excitation signal based upon the saliency of the rotor
and the location of the sensor coil. The two sensor coils are located 90 electrical degrees
apart from each other. Their output is a rotor position amplitude modulated signal as in
Figure 41.

Figure 41 Amplitude modulated rotor position signals.
A high frequency signal is injected into the stator and this signal is modulated by the
saliency of the rotor. The middle waveform is the line to line back emf of phase “a” to
“b”.
The upper waveform and lower waveform are the modulated signals where the high
frequency signal is not discernible (appears opaque) in the above image due to the
74

frequency and time scale being used. The time scale was set so that the low frequency
modulated part of the waveform could be seen. Also notice in the image, the middle
waveform is the line to line back emf voltage of phase “a” to “b”. Shown in Figure 42 is
the resultant sin(θ) and cos(θ) waveforms which are the result of applying trigonometric
identities to the waveforms in the DSP and then applying a software low pass filter.

Figure 42 Amplitude modulated rotor position signals
(top two signals) and the resultant cos(θ) and sin(θ) rotor position signals (bottom two
signals) where θ represents the rotor position.

The lower two sinusoids are shown here prior to being scaled for the DSP so that the
positive sinusoid peak = 3.3V and the lower peak = 0V. Once the signals were properly
scaled and phase shifted (due to the software filter), there is no need to find the value of θ
75

since the Park transform uses sin(θ) and cos(θ) as elements of its matrix. With this being
the case, it is very convenient to simply input the above signals directly into the Park
transform matrix. Once the motor was operating correctly, operating points could be
achieved and the power loss of the inverter could be measured. For each motor operating
point to be measured, a high voltage DC-Link and a low voltage DC-Link measurement
were taken. A Yokogawa WT-1600 Power meter was used to measure the voltages,
currents, and calculate the power and efficiency of the inverter. Below in Figure 46 are
the results of these operating point’s measured losses and calculated losses.

Figure 43 Chassis Dyno. used to obtain experimental results.

76

Representation
of ICE &
Rectifier
Generator

Inverter

Traction
Motor

Variac
Grid

Open

Rectifier and Inverter
Variac

Figure 44 Experimental Setup

77

DSP Board
and Signal
Conditioning
Circuit
Boards

Signal
Generator

Internal
Permanent
Magnet
AC Machine

Figure 45 More of the experimental setup.
These items are located below the vehicle that is on the chassis dyno.

Vehicle velocity

2.6 MPH

98

Vdc

440

5.0 MPH

197

444

7.7 MPH

292

431

10.0 MPH

305

427

Pdc

1,911 2,330 3,558 3,398 5,354 5,838 5,382 5,871

Pac

1,751 1,808 3,294 2,919 5,076 5,386 5,099 5,470

Ploss (Measured)

160

433

264

479

278

452

283

401

Ploss (Calculated)

155

420

237

430

321

436

311

407

Figure 46 Experimental results showing four operating points
each with a high voltage result and a low voltage result. Actual losses and calculated
losses are also shown.

78

Based upon these results, it can be seen that the high voltage DC-Link case does produce
higher losses in the traction inverter. Also, the measured and calculated losses show
good correspondence which lends credibility to the equations being used to calculate the
losses of the new topologies to be analyzed in the next section.

5.1.3 Base Line Continued
The total range of operating points for the hybrid vehicle is very wide. The bus speed
can range from 0 mph to 75 mph and the traction motor torque can range from a large
negative value to a large positive value for all bus speeds. Also, the choice can be made
to draw power from the ICE and generator, from the battery pack, or some combination
of the two. The input to the hybrid vehicle’s operating point decision model are gas
pedal or brake pedal position, bus speed, and battery state of charge (SOC). For any
given bus driver pedal position input, which corresponds to a requested positive or
negative traction motor(TM) torque, there is an infinite combination of generator and or
battery power levels from which to choose from. For this analysis, a reasonable range of
such operating points was sought so that a fair comparison could be made.

To illustrate a representative set of operating points, the traction motor power was swept
from 16.7kW to 250kW while holding the traction motors at full current. This represents
full acceleration which is normal for these heavy vehicles. Full acceleration, or wide open
throttle (WOT), for the typical automobile yields approximately a 0.5g acceleration.

79

Given that the typical automobile has a powertrain capable of 110kW and the bus
powertrain is approximately twice this value, yet has 10 times more mass, we have:

F  ma

(5.14)

F
a
m

(5.15)

Fa
 aa  0.5( g )
ma

(5.16)

F b 2Fa

(5.17)

mb  10ma

(5.18)

2 Fa
 ab  0.1( g )
10 ma

(5.19)

Where:
F = force
m = mass
a = acceleration
Fa = Force of an automobile at WOT
ma = mass of an automobile
aa = acceleration of an automobile
Fb = force of a bus
mb = mass of a bus
ab = acceleration of a bus
80

As shown above, it can be seen that WOT acceleration for the bus is a rather mild
acceleration of 0.1g. Additional analysis was done in this work at throttle positions other
than WOT and as can be deduced after further reading, very similar results were
observed.

For the sweep of TM operating points, three different scenarios will be shown:
Scenario 1 holds the generator power steady at 50kW throughout the TM power
sweep while the battery pack provides or absorbs the difference in power
between the generator power and TM power;
Scenario 2 holds the generator power steady at 100kW while, again, the battery
makes up the difference, and
Scenario 3 holds the generator power steady at 150kW and again, the battery
pack makes up the difference.
Figure 47shows this power sweep for the traditional hybrid topology.

81

Traditional Series Hybrid Bus:
Inverter, Converter & Inductor Loss
Power Loss (kW)

15.0
13.0
11.0

Traditional; Gen
Pwr=150kW

9.0
7.0

Traditional; Gen
Pwr=100kW

5.0
16.7
33.3
50.0
66.7
83.3
100.0
116.7
133.3
150.0
166.7
183.3
200.0
216.7
233.3
250.0

3.0

Traditional; Gen
Pwr=50kW

Traction Motor Power (kW)

Figure 47. Power loss curves for the traditional back to back VSI/VSI
series hybrid bus.
Now that the base line data has been completed, the proposed topologies can be
investigated.

5.2 Proposed Topologies
Several new topologies are proposed below. Each topology is an attempt to decouple the
DC voltage at the GI end of the topology from the DC voltage at the TI end. The power
losses of the following topologies have been calculated in the same manner as the base
line except that the voltage at the TI can be different from the GI voltage. Also, any
additional converter and their related switching states and or components (inductors for
example) are also included in the loss calculations.
assumed here as well.
82

The same three scenarios are

5.2.1 VSI / VSI with Conventional Mid-Link DC / DC Boost Converters

G

M

Figure 48 Back to back VSI to VSI with boost converters in the DC-Link

Power Loss (kW)

Inverter, Converter & Inductor Loss
14.0
12.0
10.0
8.0
6.0
4.0

Traditional; Gen
Pwr=50kW
Buck/Boost in DC-Link;
Gen Pwr=50kW
Traction Motor Power (kW)
(a) Scenario 1. Generator held at 50kW.

Figure 49 Loss graphs (a, b, and c) for the Back to back VSI to VSI with
boost converters in the DC-Link

83

Figure 49 continued

Power Loss (kW)

Inverter, Converter & Inductor Loss

14.0
9.0

Traditional; Gen
Pwr=100kW

4.0

Buck/Boost in DC-Link;
Gen Pwr=100kW
Traction Motor Power (kW)
(b) Scenario 2. Generator held at 100kW.

Power Loss (kW)

Inverter, Converter & Inductor Loss

14.0
9.0

Traditional; Gen
Pwr=150kW

4.0

Buck/Boost in DC-Link;
Gen Pwr=150kW
Traction Motor Power (kW)
(c) Scenario 3. Generator held at 150kW.

84

5.2.2 VSI / VSI Structure, LV Battery with Boost Converter, & DCLink with Buck Converter

(+)

G

M

(-)
Figure 50 Back to back VSI / VSI topology with battery boost converter and TI DCLink buck converter.

Inverter, Converter & Inductor Loss

Power Loss (kW)

14.00

12.00
10.00

Traditional; Gen
Pwr=50kW

8.00

Buck in DC-Link; Gen
Pwr=50kW

6.00
4.00
17

50 83 117 150 183 217 250
Traction Motor Power (kW)
(a) Scenario 1. Generator held at 50kW.

Figure 51 Loss graphs (a, b and c) for the Back to back VSI / VSI topology
with battery boost converter and TI DC-Link buck converter.

85

Figure 51 continued
Inverter, Converter & Inductor Loss

Power Loss (kW)

14.00
12.00
10.00
8.00

Traditional; Gen
Pwr=100kW

6.00

Buck in DC-Link; Gen
Pwr=100kW

4.00
17

50 83 117 150 183 217 250
Traction Motor Power (kW)
(b) Scenario 2. Generator held at 100kW.

Inverter, Converter & Inductor Loss

Power Loss (kW)

14.00
12.00
10.00
8.00

Traditional; Gen
Pwr=150kW

6.00

Buck in DC-Link; Gen
Pwr=150kW

4.00
17

50 83 117 150 183 217 250
Traction Motor Power (kW)
(c) Scenario 3. Generator held at 150kW.

86

5.2.3 VSI / VSI Structure, LV Battery with Buck-Boost Converter, &
DC-Link with Buck Converter

(+)

(+)

M

S4

(-)

VDC_TI

S3

VDC_GI

G

(-)

Figure 52 Back to back VSI / VSI topology with battery buck-boost converter
and GI DC-Link buck converter.

Power Loss (kW)

Inverter, Converter & Inductor Loss
18.0
16.0
14.0
12.0
10.0
8.0
6.0
4.0

Traditional; Gen
Pwr=50kW
Buck/Boost in DC-Link;
Gen Pwr=50kW

Traction Motor Power (kW)

Figure 53 Loss graphs under scenario one conditions (Generator at 50kW)
for the Back to back VSI / VSI topology with battery buck-boost converter and GI DCLink buck converter.
87

Since this topology does not provide significant savings, only the 50kW case is
displayed.

5.2.4 VSI / VSI Structure, LV Battery, DC-Link with Bi-Directional
Buck and Boost Converters

G

M

Figure 54 Back to back VSI / VSI topology with GI bi-directional buck-boost
converter and TI bi-directional buck-boost converter.

Inverter, Converter & Inductor Loss

Power Loss (kW)

14.0
12.0
10.0

Traditional; Gen
Pwr=50kW

8.0

Buck/Boost in DC-Link;
Gen Pwr=50kW

6.0
4.0

Traction Motor Power (kW)
Figure 55 Loss graphs under scenario one conditions (Generator at 50kW)
for the Back to back VSI / VSI topology with GI bi-directional buck-boost converter
and TI bi-directional buck-boost converter
88

Since this topology does not provide significant savings, only the 50kW case is
displayed.

5.2.5 Back to Back Voltage Fed Quasi-Z-Source Inverters, LV Battery
G

M

Figure 56 Back to back voltage fed quasi Z source inverter topology.

Inverter, Converter & Inductor Losses

Power Loss (kW)

14.0
12.0
10.0

Traditiona;l Gen
Pwr=50kW

8.0

VFqZSI; Gen
Pwr=50kW
16.7
33.3
50.0
66.7
83.3
100.0
116.7
133.3
150.0
166.7
183.3
200.0
216.7
233.3
250.0

6.0

Traction Motor Power (kW)
(a) Scenario 1. Generator held at 50kW.
Figure 57 Loss graphs (a, b, and c) for the Back to back voltage fed quasi Z source
inverter topology

89

Figure 57 continued.
Inverter, Converter & Inductor Losses

Power Loss (kW)

14.0
12.0
10.0

Traditional; Gen
Pwr=100kW

8.0

CF-back to back qZSI;
Gen Pwr=100kW

6.0

Traction Motor Power (kW)
(b) Scenario 2. Generator held at 100kW.
Inverter, Converter & Inductor Losses

Power Loss (kW)

14.0
12.0
Traditional; Gen
Pwr=150kW

10.0

CF-back to back qZSI;
Gen Pwr=150kW

8.0
6.0

(c) Scenario 3. Generator held at 150kW.

90

5.2.6 Back to Back Current Fed Quasi-Z-Source Inverter, LV Battery

G

M

VDC_TI

D2

D1
S1

VBatt

VDC_GI

S2

= RB-IGBT
Figure 58 Back to back current fed quasi Z source inverter topology.

Traditional topology vs. Proposed CF-qZSI topology:
Inverter, Converter & Inductor Power Loss

Power Loss (kW)

15.0
13.0
11.0
9.0

Traditional; Generator
Pwr=50kW

7.0

CFqZSI; Generator Pwr
= 50kW

5.0
3.0

Traction Motor Power (kW)
(a)

Scenario 1. Generator held at 50kW.

Figure 59. Power loss comparisons between the traditional topology and
the proposed CF-qZSI topology under three different scenarios (a), (b), (c).

91

Figure 59 continued

Traditional topology vs. Proposed CF-qZSI topology:
Inverter, Converter & Inductor Power Loss

Power Loss (kW)

15.0
13.0
11.0
9.0

Traditional; Generator
Pwr=100kW

7.0

CFqZSI; Generator
Pwr=100kW

5.0
3.0

Traction Motor Power (kW)
(b)

Scenario 2. Generator held at 100kW.

Traditional topology vs. Proposed CF-qZSI topology:
Inverter, Converter & Inductor Power Loss

Power Loss (kW)

15.0
13.0

11.0
9.0

Traditional; Generator
Pwr=150kW

7.0

CFqZSI; Generator
Pwr=150kW

5.0
3.0

Traction Motor Power (kW)
(c)

Scenario 3. Generator held at 150kW.
92

Although it can be reasoned that the additional components reduce system reliability, the
Quasi Z Source Inverter topologies have the added quality that they are not vulnerable to
shoot through events which would destroy the switching devices of the traditional VSI
topologies [16]. The above topologies are based upon various types of buck, or boost, or
buck boost converters implemented in a way that decouples the DC-Link voltage at one
end from that at the other. The most promising of these topologies is the back to back
current fed qZSI (CF-qZSI) shown in Figure 58. The CF-qZSI will be discussed here in
further detail. First it is necessary to discuss the functionality of the CF-qZSI.

Current Fed Quasi “Z” Source Inverter (CF-qZSI):
The CF-qZSI uses reverse blocking IGBTs (RB-IGBT) without a free-wheeling diode.
This device is able to block voltage in both directions, but can pass current in only one
direction. It has slightly higher forward conduction losses, but does not have the freewheeling diode or the losses that are associated with it. Like the other “Z” source type
inverters, the inverter switching devices include an additional zero switching state that is
not allowed with the traditional inverters. This additional switching state is frequently
referred to as the shoot-through state or short circuit zero state when both the top and
bottom switches of a phase leg are closed at the same time. This switching state would
destroy the switching devices of the traditional inverter but the switching currents are
safely limited in the “Z” source inverters by the inductance of these inverters.
The operation and switching states used to control the circuit configuration at the traction
motor end (from the battery to the traction motors) is different than that at the generator
93

end (from the battery to the generator). Therefore, each will be discussed separately
below:
Traction Motor End:
The operation of the traction motor end of the back to back topology is shown in Figure
60. Figure 60 (a), (b), and (c) show the CF-qZSI active, shoot-through (short circuit zero
state), and open (open circuit zero) states, respectively. Figure 61 shows the voltage
boost or buck results as a function of these states.

LT3
SWu
LT1

+
CT2
-

CT1

+
V
- out
LT2

SWL

(a) State I– Active State: DA

LT3
LT1

CT1

SWu
+
CT2
-

+
Vout
LT2

SWL

(b) State II– Short Zero State: Dsh
Figure 60. Operational states (a, b, and c) of the CF-qZSI [27]

94

Figure 60 continued

LT1

CT1

+
CT2
-

+
Vout
LT2

-

(c) State III– Open Zero State: DOP

When the inverter uses only active states and not zero states, the DC output voltage is
equal to the DC input voltage. If boosting this voltage is desired, the active duty cycle
will be reduced and short circuit zero states will be added. If bucking is desired, the
active duty cycle will be reduced and open circuit zero states will be added. This does
not reduce the DC voltage across the TI below the battery voltage, but reduces the
voltage seen by the traction motor. In the graph of Figure 61, the red line (upper
boundary) is purely active states and short circuit zero states while the blue line (lower
boundary) is purely active states and open circuit zero states.
boundary lines, a combination of all three states can be used.

95

Between these two

5
4
3

Region A

2

Vout
Vin 1
0

Mode 1

Region B
Region C

-1

Mode 2

-2
-3

0.2 0.3 0.4 0.5 0.6

0.7 0.8

0.9 1.0

DA
Figure 61. Vout/Vin vs. active duty cycle. [27]
Mode 1 uses active states and short circuit zero states but not open circuit zero states,
Mode 2 uses active states and open circuit zero states but no short circuit zero states.

At the traction inverter end of the topology, it is this buck and boost capability that allows
the energy saving operation of maintaining the DC-Link voltage across the traction
inverters to be as low as the battery voltage and as high as is required for full speed
operation of the bus.

96

Generator End:
The operation of the qZSI configuration at the generator end of the topology is a bit
different. The qZSI at this end is used to buck the generator voltage down to the level of
the battery by using the active and open circuit zero states.

Vbatt
VDC GI

 D A (eq 5.20)

Where:

DA is the active duty cycle of the generator inverter switches,
Vbatt is the battery voltage, and
VDC-GI is the DC-Link voltage across the generator inverter.

In this case, as the active duty cycle is reduced from 100%, it is the open circuit zero state
that is increased and the short circuit zero state is not used while pushing power from the
generator to the battery and traction inverter. Figure 62 shows the generator end of the
topology.

97

VDC_GI

G
D1

S1

Figure 62. Generator end of the back to back CF-qZSI topology

This end of the topology is not a complete mirror image of that at the traction motor end.
The RB-IGBTs are flipped over allowing the top DC-Link rail to be polarized with a
positive voltage from the generator and an additional IGBT has been added in the middle
of the “Z” source impedance network. This switch “S1” remains open during normal
operation and is only closed when power is to be pushed from the battery to the generator
to start the diesel engine. Also, there is an additional switch S2 which is gated on at all
times except when the ICE is being started. When gated on, it functions like a diode and
conducts in the state shown in Figure 63.

While the generator is running and power is being pushed from the generator to the
battery and the TI, the generator’s voltage level must be bucked. To accomplish this, the
following switching states are used.
98

+
-CG2

LG1

LG3

Vout
+
-

LG2

Vin
CG1

(a) GI active state with power being pushed to the battery and TI. S1 and S2 are open

+
C
- G2

LG1

LG3

Vout
S2
LG2

+
-

CG1

(b) GI open circuit zero state with S2 conducting. S2 is closed, S1 is open.

Figure 63. (a) and (b) GI voltage is being bucked from the GI level
down to the battery level.

During state 1 shown in Figure 63 (a), generator voltage is imposed across inductor LG3
and the battery. Inductor LG3 provides a voltage drop and the battery only sees a voltage
lower the VDC_GI. The duty cycle of this state is controlled to maintain the correct
voltage across the battery. Following this state, the state of Figure 63 (b) is entered.
During this state, switch S2 turns on and allows the currents in the inductors to decay.
Below in Figure 64, the switching states used to start the diesel ICE are displayed. Since
the GI uses RB-IGBTs, power cannot be pushed toward the generator by simply
99

reversing the flow of current as can be done with the more common VSI. In this case,
voltage polarity must be reversed to reverse power flow. Figure 64 (a) shows the initial
state. In this state, S1 is closed which impresses the voltage of CG2 across LG1, and CG1
across LG2. The voltage across each capacitor is equal to the battery voltage. This
energizes these inductors in preparation for the next state. In Figure 64 (b), S1 opens and
the GI enters its active state. This pushes power to the generator and the GI now
functions to drive the generator as a starter motor.

+
C
- G2

LG3

LG1

Vout
S1 +
LG2
-

CG1

(a) Open circuit zero state with S1 closed and S2 open.

+
-CG2

LG1

LG3

Vout
LG2

+
-

Vin
CG1

(b) Active state with S1 and S2 open.
Figure 64. The switching states used to start the diesel ICE

100

Calculation of Losses for the proposed back to back CF-qZSI topology:
The power losses for the CF-qZSI were calculated in a manner similar to that used for
the traditional hybrid bus of figure 3. However, there are several differences that must be
taken into account. The CF-qZSI uses reverse blocking IGBTs (RB-IGBT) instead of the
common IGBT module with the anti-parallel, free-wheeling diode. The RB-IGBT has
slightly higher forward conduction losses, but does not have the losses from an antiparallel free-wheeling diode. The qZSI impedance network on the generator inverter side
includes an IGBT module (IGBT and its anti-parallel diode) which does not exist in the
traditional hybrid topology and the switching, conduction, and reverse recovery losses of
this device have to be considered for the modes under which they turn on, off or conduct
accordingly. You will also notice the two inductors of this impedance network and an
inductor connected to the battery node. The losses of these inductors have also been
estimated in the same manner as that of eq. 5.13 based upon the power flow through them
and the core losses (hysteresis and eddy) associated with the changing magnetic field
(delta B) which results from boost or buck functions. Also, the capacitors of the qZSI
impedance network are equivalent to those of the traditional topology and the
comparative losses were found to be inconsequential and therefore, the capacitor losses
of both the traditional and proposed topologies are ignored in this study.

The inductor losses on the inductor on the TM side of the battery node and of the
impedance network for the TM qZSI are likewise calculated as well as the conduction
and reverse recovery losses of the diode in that impedance network. Also, the Traction
inverter has an additional switching state that the traditional VSI does not have [16] [27].
101

This shoot-through state in which both the top and bottom switching devices of a phase
leg are gated on at the same time, provides an important function of this unique type of
inverter that allows it to boost the voltage seen at the input to the traction inverter, and the
losses of this additional switching state (turn on, turn off, and conduction) are also
included in the analysis.

The Z-source topology does not require the dc/dc converter and therefore avoids the
switching and conduction losses associated with that; however, it does utilize an extra
switching state for shoot-through. This state occurs during the open circuit zero state
during which all 6 RB-IGBTs are open. The OCZ state of the CF-qZSI is different than
that of the voltage fed type of inverters where all 3 upper IGBTs are closed or all 3 lower
IGBTs are closed to create this state. When the TI of the CF-qZSI enters the SCZ state, it
closes the upper and lower switches of one phase leg leaving the other four RB-IGBTs
open. Since the RB-IGBTs conducting the SCZ current had been open, and will return to
of the open state, these two switches incur a turn on, turn off, and conduction loss
associated with this SCZ state. This also is included in the analysis, and the following
power loss equations are added to the Z-source topology.

PON _ ST

Pi _ ST

Vcc Mt rN I cM 2 Fs cos( )

2I cN

 V  V
c0
   ceN

I cN



 1
 MI cos( )   V  
 c0

 2 cM



102

(eq 5.21)

1
T  M
 MI cM cos( )  1 
B
2
2

Poff _ ST


 Fs


(eq 5.22)

 Vcc MI cM t fN Fs cos( )  MI cM cos( ) 
1 


 (eq 5.23)


6
4 I cN




Where the variables are as defined at eq 5.4.

To summarize the loss effects of the traditional hybrid vs. the proposed back to back CFqZSI:
o The proposed topology has additional losses that result from
additional components and switching states:


“Z” source impedance network inductors and Inductors
connected to the battery node



Reverse blocking feature of the inverter switching devices



Additional IGBT module included in the generator end “Z”
source network



Additional diode included in the traction motor end “Z”
source impedance network, and



Additional switching states of the “Z” source type
inverters.

o The proposed topology has lower losses that result from:


not having a separate DC/DC converter connected to the
battery, and



The new ability to decouple the DC-Link voltage present at
opposite ends of the topology.

103

All of the positive and negative loss effects of both the traditional hybrid topology and
the proposed topology were then aggregated for comparison as shown in Figure 59.

5.3 Conclusion
Several topologies were devised analyzed and presented with the purpose of seeking a
topology that would allow each major component of the hybrid bus to operate within its
most efficient region. Three topologies in particular do show significant decreases in
losses for lower speed operation of the bus which is its most common range of speeds.
These new more efficient topologies are the:
(1) VSI / VSI Structure with a LV Battery and Boost Converter, & a DC-Link
with a Buck Converter;
(2) Back to Back Voltage Fed Quasi-Z-Source Inverters with a LV Battery; and
(3) Back to Back Current Fed Quasi-Z-Source Inverters with a LV Battery.
The commuter bus optimized with one of these new topologies would be an ideal
candidate for implementation as a parallel or series parallel hybrid.

In such an

implementation, the additional high power operation could be accommodated via the
parallel mechanical power flow path. This would allow the power electronics to always
remain within the more efficient power and speed ranges. In effect, this would utilize the
electric components to function in the regions where they provide more efficiency than
the traditional hybrid vehicle and allow a mechanical power flow path to deliver just the
high power in the regions of operation where it is the most efficient.
104

Of these three advanced topologies, the “VSI / VSI Structure with a LV Battery and
Boost Converter, & DC-Link with a Buck Converter” (Here after referred to as “VSI/VSI
with DC-Link Buck”) topology and the “Back to Back Current Fed Quasi-Z-Source
Inverters with a LV Battery” (hereafter referred to as “Back to Back CF-qZSI”) provide
the most energy savings. The VSI/VSI with DC-Link Buck is the most simple and yet
provides significant energy savings. The Back to Back CF-qZSI offers the lowest energy
loss in a wider range of operating regions. Below in Figure 65 are comparative graphs
for these two topologies and the traditional topology. Figure 65 (a) shows the traction
motor power sweep while the generator power is held at 50kW (scenario 1). Figure 65
(b) shows the same traction motor power sweep while the generator is held at 100kW
(scenario 2), and Figure 65 (c) shows the same sweep while the generator is held at
150kW (scenario 3).

105

Inverter, Converter & Inductor Loss

Power Loss (kW)

14.0
12.0
Traditional; Gen
Pwr=50kW

10.0
8.0

DC-Link Buck; Gen
Pwr=50kW

6.0

CF-qZSI; Gen
Pwr=50kW
17
33
50
67
83
100
117
133
150
167
183
200
217
233
250

4.0
Traction Motor Power (kW)

(a) traction motor power sweep while the generator power is held at 50kW (scenario 1)

Inverter, Converter & Inductor Loss
Power Loss (kW)

14.0

12.0

CF-qZSI; Gen
Pwr=100kW

10.0
8.0

DC-Link Buck; Gen
Pwr=100kW

6.0

Traditional; Gen
Pwr=100kW

17 50 83 117 150 183 217 250
Traction Motor Power (kW)

(b) traction motor power sweep while the generator power is held at 100kW (scenario 2)

Figure 65 Comparative power loss graphs for the Traditional,
Buck Converter in DC-Link,
and CF-qZSI topologies for three different operating scenarios (a), (b), and (c).
106

Figure 65 continued

Inverter, Converter & Inductor Loss
Power Loss (kW)

14.0

12.0

Traditional; Gen
Pwr=150kW

10.0

DC-Link Buck; Gen
Pwr=150kW

8.0
6.0
17 50 83 117 150 183 217 250
Traction Motor Power (kW)

CF-qZSI; Gen
Pwr=150kW

(c) traction motor power sweep while the generator power is held at 150kW (scenario 3)

As shown in the graphs above, the new topologies offer significant savings of energy at
important operating points. A summary of the pros and cons of the VSI/VSI with DCLink Buck and the Back to Back CF-qZSI is presented just below.

107

VSI/VSI with DC-Link Buck:
Pros:


decreased losses for traction motor power below 117kW.



an additional inductor.



two additional IGBT modules.

Cons:

Back to back CF-qZSI:
Pros:


lowest losses at low vehicle speed.



lowest losses over a wide speed range for scenario 1.



system is protected against shoot through events.



freewheeling diode count is very low.



additional inductors.



two additional IGBT modules.

Cons:

As a result of this work, a new full system analysis has been done, and based upon this
analysis, a new generator inverter to traction inverter voltage decoupled structure has
been proposed.

The voltage decoupling method has been proven via experimental
108

results, and the analytical method implemented in this work has been verified via
experimental results. Several new voltage decoupling topologies have been developed
and analyzed. The analysis method has been used to choose the two best topologies
(VSI/VSI with DC-Link Buck, and Back to Back CF-qZSI) which have been compared to
each other and to the existing traditional topology.

Significantly improved efficiencies

have been achieved by this research and the new topologies proposed as its result.

109

6. Future Work

Future work recommend is to devise a by-pass method that allows the
lower high speed power losses of the traditional topology as well as the lower low speed
power losses of the new topologies. A type of parallel hybrid or series parallel hybrid
could be a good candidate for this. Also, a type of electrical bypass solution should be
sought that allows the voltage decoupling systems to be by-passed when they are not
needed.

As a result of the voltage decoupling concept, the traction inverters are allowed to operate
at a modulation index of one and it is known that this reduces the harmonic content seen
by the traction motors thereby improving their efficiency as well. The extent to which
this occurs should be quantified.

The Back to Back CF-qZSI topology has the capability to boost the DC-Link voltage. As
a result, the traction motors could be kept below the field weakening region for a longer
period of time by increasing the output voltage of the inverters as the field weakening
region is approached. This allows the additional current required to weaken the rotor
field to be avoided and thereby reduce losses of the traction motors. This should also be
quantified.

110

Thirdly, the new topologies allow the speed of the diesel engine and therefore the speed
and voltage of the generator to be held in a narrower range. As a result, the efficiency of
operation of the engine and of the generator could be optimized for this operating point
and result in greater energy savings which should also be quantified.

The power electronics energy savings of the new topologies have been investigated in
this work, but there are additional savings that result from the whole system approach
used here. As suggested in the previous three paragraphs, the new topologies provide
additional savings opportunities in the diesel engine, generator, and traction motors that
have not yet been quantified. This future work will show that the new topologies are
even more beneficial than documented here.

111

REFERENCES

112

REFERENCES

[1] Bellis, Mary, “The History of the Automobile – Early Steam Powered Cars,”
October 14, 2007, http://inventors.about.com/library/weekly/aacarssteama.htm
[2] New World Encyclopedia , “Internal Combustion Engine”, May 05, 2009,
http://www.newworldencyclopedia.org/entry/Internal_combustion_engine
[3] M.J. Bradley, and Associates; Report 59 Hybrid-Electric Transit Buses: Status, Issues,
and Benefits. Washington, D.C.: National Academy Press, 2000.
[4] H. Moghbelli, K. Ganapavarapu, and R. Langari, “A comparative review of fuel cell
vehicles (FCVs) and hybrid electric vehicles (HEVs) – Part II,” in Proc. SAE Future
Transportation Technology Conf., Costa Mesa, CA, June 2003.
[5] Syed, Fazal U.; Kuang, Ming L.; et. al. “Derivation and Experimental Validation of a
Power-Split Hybrid Electric Vehicle Model.” November 2006, IEEE Transactions on
Vehiclular Technology, Vol 55, No. 6
[6] R. Thring; Fuel Cells for Automotive Applications. New York, NY: The Americal
Society of Mechanical Engineers, 2004.
[7] K. Thorborg; Power Electronics. London, U.K.: Prentice-Hall International (U.K.)
Ltd., 1988.
[8] A. Emadi, S. S. Williamson; “Fuel Cell Vehicles: Opportunities and Challenges,”
Grainger Power Electronics and Motor Drives Laboratory, Electric Power and Power
Electronics
Center,
Illinois
http://www.ece.iit.edu/~emadi/

Institute

of

Technology,

Chicago,

IL,

[9] M.H. Rashid; Power Electronics, 2nd ed. Englewood Cliffs, NJ: Prentice-Hall, 1993.

113

[10] R. Erickson, D. Maksimovic; Fundamentals of Power Electronics, 2nd ed. Norwell,
MA: Kluwer Academic Publishers, 2001.
[11] G. Ledwich, “Pulse Width Modulation (PWM) Basics,” website, 1998.
http://www.powerdesigners.com/InfoWeb/design_center/articles/PWM/pwm.shtm
[12] V. Onde; “ ST92141 AC Motor Control Software Library Version 2.1 Update,”
AN1367 Application Note, STMicroelectronics, 2002.
[13] D.A. Grant and J. A. Houldsworth: PWM AC Motor Drive Employing Ultrasonic
Carrier. IEE Conf. PE-VSD, London, 1984, pp. 234-240.
[14] V. Onde; “ ST92141 AC Motor Control Software Library Version 2.1 Update,”
AN1367 Application Note, STMicroelectronics, 2002.
[15] D.A. Grant and J. A. Houldsworth: PWM AC Motor Drive Employing Ultrasonic
Carrier. IEE Conf. PE-VSD, London, 1984, pp. 234-240.
[16] F. Z. Peng, “Z-Source Inverter,” IEEE Transactions on Industry Applications, vol.
39, No. 2, pp. 504-510, March/April 2003.
[17] F. Z. Peng and Miaosen Shen, Zhaoming Qian, “Maximum Boost Control of the Zsource Inverter,” in Proc. of IEEE PESC 2004.
[18] Miaosen Shen, Jin Wang, Alan Joseph, Fang Z. Peng, Leon M. Tolbert, and Donald
J. Adams, “Maximum constant boost control of the Z source inverter.” in Proc. IEEE
IAS’04, 2004, p.142.
[19] Srinivas Iyer, “The Mean Time-to-Failure of some Special Series-Parallel
Arrays”,IEEE Transactions on Reliability, Vol. 40, No. 1, 1991 April
[20] Ruifang Liu, Chris Chunting Mi, David Wenzhong Gao, “Modeling of Iron Losses
of Electrical machines and Transformers Fed by PWM Inverters”, Power Engineering
Society General Meeting, 2007, IEEE 24-28 June 2007 Pages: 1-7, digital object
identifier 10.1109/PES.2007.385873
114

[21] K. D. Holland; “Z-Source Inverter Control for Traction Drive of Fuel Cell – Battery
Hybrid Vehicles,” Thesis presented to Michigan State University, Department of
Electrical and Computer Engineering, August 26, 2005

[22] Mohan, Ned; Underland, Tore M.; Robbins, William P.; “Power Electronics”, 2003
John Wiley & Son’s, Inc. , Hoboken, New Jersey
[23] Sze, S.M.; Kwok, K. NG; “Physics of Semiconductor Devices”, 2007 John Wiley &
Sons, Inc., Hoboken, New Jersey
[24] Moriya, K.; Nakai, H.; Inaguma, Y.; Ohtani, H.; Sasaki, S.;”A novel multifunctional converter system equipped with input voltage regulation and current ripple
suppression”; Industry Applications Conference, 2005. Fortieth IAS Annual Meeting.
Conference Record of the 2005 , Page(s): 1636 - 1642 Vol. 3
[25] Casanellas, F., “Losses in PWM iverters using IGBTs”; IEE Proc.-Electr. Power
Appl., Vol 141, No. 5, September 1994
[26] Xi Lu; Rogers, C.; Fang Zheng Peng; , "A double Fourier analysis development of
THD for PWM inverters: A theoretical method for motor loss minimization," Energy
Conversion Congress and Exposition (ECCE), 2010 IEEE , vol., no., pp.1505-1510, 1216 Sept. 2010
[27] Qin Lei; Shuitao Yang; Fang Zheng Peng; Inoshita, R.; , "Application of current-fed
quasi-Z-Source Inverter for traction drive of hybrid electric vehicles," Vehicle Power and
Propulsion Conference, 2009. VPPC '09. IEEE , vol., no., pp.754-760, 7-10 Sept. 2009
[28] U. S. Department of Energy; “Energy Efficiency and Renewable Energy – MultiYear Program Plan 2011-2015 Vehicle Technologies Program”; December 2010, p. 2.3-2

115