Michigan Technological University

Digital Commons @ Michigan Tech

Dissertations, Master's Theses and Master's Dissertations, Master's Theses and Master's
Reports - Open Reports

2013

MODELING AND HARDWARE-IN-THE-LOOP SIMULATION OF

POWER-SPLIT HYBRID ELECTRIC VEHICLES
Shreyash Joshi
Michigan Technological University

Follow this and additional works at: https://digitalcommons.mtu.edu/etds

Part of the Mechanical Engineering Commons

Copyright 2013 Shreyash Joshi

Recommended Citation
Joshi, Shreyash, "MODELING AND HARDWARE-IN-THE-LOOP SIMULATION OF POWER-SPLIT HYBRID
ELECTRIC VEHICLES", Master's report, Michigan Technological University, 2013.
https://digitalcommons.mtu.edu/etds/824

Follow this and additional works at: https://digitalcommons.mtu.edu/etds

Part of the Mechanical Engineering Commons
MODELING AND HARDWARE-IN-THE-LOOP SIMULATION OF POWER-SPLIT HYBRID
ELECTRIC VEHICLES

By

Shreyash Joshi

A REPORT

Submitted in partial fulfillment of the requirements for the degree of

MASTER OF SCIENCE

In Mechanical Engineering

MICHIGAN TECHNOLOGICAL UNIVERSITY

2013

© 2013 Shreyash Joshi

This report has been approved in partial fulfillment of the requirements for the
Degree of MASTER OF SCIENCE in Mechanical Engineering.

Department of Mechanical Engineering-Engineering Mechanics

Report Advisor: Dr. Bo Chen

Committee Member: Dr. John Beard

Committee Member: Dr. Zhuo Feng

Department Chair: Dr. William W. Predebon

TABLE OF CONTENTS

LIST OF FIGURES ......................................................................................... V

LIST OF TABLES ........................................................................................ VII

ACKNOWLEDGEMENTS ............................................................................. VIII

ABSTRACT ................................................................................................... IX

1. INTRODUCTION ................................................................................. 10

2. LITERATURE REVIEW ......................................................................... 16

2.1. POWER-SPLIT HYBRID ELECTRIC VEHICLE CONFIGURATIONS .............................. 16

2.2. HEV CONTROL STRATEGIES .................................................................... 17

3. DYNAMIC MODEL OF POWER-SPLIT HEV POWERTRAIN ..................... 18

3.1. SINGLE PLANETARY GEAR SET ........................................................... 18

3.2. HYBRID SYNERGY DRIVE ................................................................... 19
3.3. DYNAMICS OF THS III GEAR TRAIN (HYBRID SYNERGY DRIVE) .......................... 23
3.4. MODIFIED HYBRID SYNERGY DRIVE............................................................ 25
3.5. DYNAMIC MODEL OF HYBRID SYNERGY DRIVE WITH CLUTCH .............................. 26
3.5.1. Electric Only mode (EV) .........................................................................................................26
3.5.2. Hybrid Mode ..........................................................................................................................29

4. HIL SIMULATION VEHICLE MODEL ..................................................... 33

4.1. OVERVIEW OF VEHICLE MODEL ................................................................. 33

4.2. VEHICLE ECU NETWORK ........................................................................ 35
4.3. VEHICLE PLANT MODEL .......................................................................... 39
4.3.1. Engine Model .........................................................................................................................39
4.3.2. Model of Electronic Components ...........................................................................................40
4.3.3. Development of Power-split Drivetrain Model ......................................................................45
4.3.4. Vehicle Dynamics ...................................................................................................................48
4.4. SENSORS AND ACTUATORS ...................................................................... 50
4.5. ENVIRONMENT .................................................................................... 50

5. HIL SIMULATION AND RESULT DISCUSSION ..................................... 51

5.1. DSPACE HIL SIMULATOR....................................................................... 51

III
 5.2. HIL SIMULATION SETUP ......................................................................... 52
5.3. HIL SIMULATION RESULTS AND DISCUSSION ................................................ 54

5. CONCLUSIONS AND FUTURE WORK ................................................... 65

REFERENCES .............................................................................................. 66

IV
LIST OF FIGURES

FIGURE 1.1: PROJECTED DEMAND AND SUPPLY OF OIL RESERVES [1]. ................ 10

FIGURE 1.2: CO2 EMISSIONS FROM VARIOUS SOURCES [2]................................ 11

FIGURE 1.3: CHANGE IN TEMPERATURE SINCE INTRODUCTION OF ICE [1]........... 12

FIGURE 1.4 : CONVENTIONAL V-MODEL ........................................................... 14

FIGURE 3.1: SINGLE PLANETARY GEAR SET ...................................................... 18

FIGURE 3.2: HYBRID SYNERGY DRIVE SYSTEM .................................................. 19

FIGURE 3.3: 3D MODEL OF SYNERGY DRIVE SYSTEM ......................................... 20

FIGURE 3.4: 3D MODEL OF COMPOUND RING GEAR ........................................... 21

FIGURE 3.5: 3D MODEL OF THS III PLANETARY GEAR SET .................................. 21

FIGURE 3.6: FREE BODY DIAGRAM OF SUN GEAR 1 (S1) AND SUN GEAR 2 (S2).... 23

FIGURE 3.7: FREE BODY DIAGRAM OF COMPOUND RING GEAR ........................... 24

FIGURE 3.8: FREE BODY DIAGRAM OF CARRIER GEAR ........................................ 24

FIGURE 3.9: MODIFIED HYBRID SYNERGY DRIVE ............................................... 25

FIGURE 3.10: POWER-SPLIT POWERTRAIN IN ELECTRIC ONLY MODE ................... 27

FIGURE 3.11: FREE BODY DIAGRAM OF MG2 ..................................................... 27

FIGURE 3.12: FREE BODY DIAGRAM OF MG1 ..................................................... 28

FIGURE 3.13: POWER-SPLIT POWERTRAIN IN HYBRID MODE .............................. 29

FIGURE 3.14: FREE BODY DIAGRAM OF ENGINE CONNECTED TO CARRIER ........... 30

FIGURE 3.15: FREE BODY DIAGRAM OF MG1 IN HYBRID MODE ........................... 30

FIGURE 4.1: OVERVIEW OF THE VEHICLE MODEL .............................................. 34

FIGURE 4.2: CONTROL STRATEGY TO SELECT VEHICLE MODE ............................. 36

FIGURE 4.3: VEHICLE ARCHITECTURE .............................................................. 39

FIGURE 4.4: DSPACE ENGINE MODEL ............................................................... 40

FIGURE 4.5: MG2 AND MG1 MOTOR EFFICIENCY MAP ......................................... 41

V
FIGURE 4.6: DSPACE ELECTRIC COMPONENTS MODEL ....................................... 42

FIGURE 4.7: POWER-SPLIT MODEL ................................................................... 45

FIGURE 4.8: VEHICLE DYNAMICS ..................................................................... 48

FIGURE 5.1: HIL SIMULATION SET-UP .............................................................. 52

FIGURE 5.2 : VEHICLE SPEED TRACKING IN UDDS............................................. 55

FIGURE 5.3: COMPARISON OF REQUESTED ICE TORQUE AND GENERATED ICE

TORQUE .................................................................................................. 56

FIGURE 5.4: COMPARISON OF BATTERY SOC WITH SOC UPPER LIMIT .................. 57

FIGURE 5.5: THE RPM OF PRIMARY MOTOR, MG2 ............................................... 58

FIGURE 5.6: THE RPM OF SECONDARY MOTOR MG1 ........................................... 59

FIGURE 5.7: COMPARISON OF ICE MODE CHANGE WITH VEHICLE VELOCITY ........ 60

FIGURE 5.8: GENERATOR RPM AND ENGINE RPM .............................................. 61

FIGURE 5.9: VEHICLE AND ICE MODE COMPARISON WITH POWERTRAIN

COMPONENTS .......................................................................................... 62

FIGURE 5.10: ENGINE OPERATING POINTS IN BSFC MAP OF 2.9 LITER ENGINE .... 63

FIGURE 5.11: MG2 MOTOR OPERATING POINTS ON MOTOR EFFICIENCY MAP ....... 64

VI
LIST OF TABLES

TABLE 1: NUMBER OF TEETH ON DRIVE GEARS ................................................. 22

TABLE 2: BATTERY SPECIFICATION .................................................................. 44

TABLE 3: CONSTANTS USED IN SIMULATION .................................................... 49

TABLE 4 : THE CONNECTIONS BETWEEN HIL SIMULATOR AND MICROAUTOBOX.... 53

VII
 ACKNOWLEDGEMENTS
I thank my advisor Dr. Bo Chen, for support in this research work and guiding me for
my Masters education. I would like to thank committee member Dr. John Beard and
Dr. Zhuo Feng for being in the committee and all their suggestions.

I am grateful to all my Intelligent Mechatronics and Embedded Systems Lab

colleagues at Michigan Technological University special thanks to Yang Li and Ming
Cheng for their help and support in my research.

I would like to thank all my friends I made at Michigan Tech. I admire your help and
support through all good and bad times.

Finally, ]_, –ȡ–ȡ ]ͨŽ ‘ȡ‘ȡ›ȡ €ȪŠȣ €ȪŠȣ ĤŽȡ˜...

VIII
ABSTRACT
Conventional vehicles are creating pollution problems, global warming and the
extinction of high density fuels. To address these problems, automotive companies
and universities are researching on hybrid electric vehicles where two different power
devices are used to propel a vehicle. This research studies the development and
testing of a dynamic model for Prius 2010 Hybrid Synergy Drive (HSD), a power-split
device. The device was modeled and integrated with a hybrid vehicle model. To add
an electric only mode for vehicle propulsion, the hybrid synergy drive was modified
by adding a clutch to carrier 1. The performance of the integrated vehicle model was
tested with UDDS drive cycle using rule-based control strategy. The dSPACE
Hardware-In-the-Loop (HIL) simulator was used for HIL simulation test. The HIL
simulation result shows that the integration of developed HSD dynamic model with a
hybrid vehicle model was successful. The HSD model was able to split power and
isolate engine speed from vehicle speed in hybrid mode.

IX
1. INTRODUCTION

The fossil fuels are being utilized as main sources of energy for most of the power
producing applications. Due to the fast depletion of oil resources, the remaining
reserves are expected to approach to zero around 2040 as shown in Figure 1.1. The
limited resources of oil also cause the raising of oil prices every year. In addition to
the loss of crucial energy resources of very high energy density, the global warming
is also a severe issue. One of the major contributors to global warming is the CO2
emissions of conventional vehicles. Figure 1.2 shows that the CO2 emission from
transportation is very close to the amount generated by industry.

Figure 1.1: Projected demand and supply of oil reserves [1].

10
 Figure 1.2: CO2 emissions from various sources [2].

The rise of CO2 emission causes the increase of global temperature. Figure 1.3 shows
temperature being steady till almost 1900. Since the inception of Internal
Combustion Engine (ICE) into vehicles and transportation, temperature is rising
steadily. With growing economy, the number of conventional automobiles sold will be
increasing. As ICE efficiency is limited to the Carnot cycle efficiency, a technological
advantage has to be gained so that the extinction of fossil reserves doesn’t hamper
the growth of humanity. To limit CO2 emissions and increase the efficiency of
vehicles, government is implementing higher CAFÉ standards which all the
automobile manufactures need to follow.

11
 Figure 1.3: Change in temperature since introduction of ICE [1].

The efficiency of conventional vehicles can only been improved to a certain extent
due to the limitation of energy conversion and heat losses. To overcome this
limitation, hybridization of the powertrain is one of the solutions. Hybrid vehicles are
powered by ICE and electric motors. There are two significant advantages of this
powertrain combination. One is the high efficiency of motor at higher torque and the
other is regenerative braking, which allows putting the mechanical energy back into
power source (battery).

Till date, research on Hybrid Electric Vehicle (HEV) is of three basic types. The first
type is series hybrid. For a series hybrid vehicle, ICE charges battery while e-motor
drives car throughout drive cycle. The second type is parallel hybrid, where either e-
motor or the ICE or both are driving the vehicle at same time. The third type is
power-split, where either e-motor/ICE can drive at once or both can work
simultaneously while charging the battery. The engine Rotation Per Minute (RPM) in
power-split vehicles is isolated from vehicle speed to achieve maximum miles/$. The
power-split allows a vehicle to act as parallel or series when needed. This comes at a
cost of added complexity to the overall system but proves to be beneficial in engine
optimization.

12
Currently, the vehicle electrification is bound by limited storage capacity of batteries
and long charging hours. Though there are some commercial manufacturing
industries like Tesla Motors Inc. which are involved in selling of complete electric
vehicle, long charging time and vehicle range anxiety scares customers away from
buying a pure electric vehicle. The power-split and plug-in hybrid electric vehicles are
best available compromise which can be implemented to lay foundation for the full
electrification plan. The plug-in hybrid electric vehicles allow charging of batteries
from the home power supply. The Volt introduced by General Motors is the prime
example of plug-in hybrid electric vehicles.

Any new powertrain design needs to be tested. The conceptual idea of hybrid vehicle
and its testing for commercial implementation could cost a lot of money to a
company. In basic vehicle designing process, firstly, the size and type of powertrain
is decided. Once the size of powertrain is decided, economy of a vehicle can be
greatly improved by implementing right control strategy. To validate whether the
implemented control strategy is optimum, a good mathematical model of all working
systems is necessary. The software-in-loop (SIL) and hardware-in-loop (HIL) testing
methods are used in testing and modification of hybrid vehicles.

The control strategy development can be implemented in two ways, with

conventional V-model as shown in Figure 1.4 [3] and model-based design. In
conventional V-model algorithm, the control specification is firstly designed. Then,
the system level and subsystem level control are designed and implemented. The
test of the control objectives are started from subsystems, system level, and then
finally integrated with vehicles. With a number of iterations, the V-model design and
test process becomes inefficient. Model-based design allows the designer and his
team to switch within any level of vehicle development avoiding time losses due to
testing and engineering hours. In model-based design, mathematical models are
implemented in graphical format which also makes debugging of algorithms very
easy.

13
 Defining Vehicle Integration &
Specification Verification

System-level System-level
Design Test

Subsystem-level Subsystem-
Design level Test

Implementation

Figure 1.4 : Conventional V-model

The mathematical models are created in coding platform of computer. The coded
system can first be validated in software platform using software-in-the-loop test. If
the performance is appropriate, the machine code of the developed models will be
generated and downloaded to Electronic Control Unit (ECU) or the HIL processor
board for HIL testing of the system. HIL test of mathematical models are in real-time
on a simulator. In this report, an ECOLine HIL simulator developed by dSPACE Inc. is
used to run simulation. The dSPACE simulator is equipped with a RTI1006 processor
board for real-time simulation. An ECU can be a hardware ECU or a softECU. One
motor ECU uses dSPACE MicroAutobox II. The other ECUs are softECUs. The
simulator is calibrated using a licensed copy of ControlDesk 7.2 software. The
ControlDesk has a user friendly Graphics User Interface (GUI). ControlDesk GUI
allows us to build different layouts and measuring panels to run and calibrate
experiments in real-time. The dSPACE package comes with two completed models of
a conventional ICE vehicle and a parallel hybrid electric vehicle. The models can be
edited in Simulink platform. Auto-code generation software converts the Simulink
models into C language code. This code can be easily compiled into machine code for
microprocessors. For this project, HIL simulator along with MicroAutobox is being
used to simulate a complete power-split hybrid vehicle in ControlDesk environment.

The report has been organized in following chapters.

Chapter 1: discusses the motivation and need for this research work.

Chapter 2: literature review of various power-split configurations and control

strategies.
14
Chapter 3: dynamic model of power-split HEV, including all the mathematical
equations needed for the development of power-split model.

Chapter 4: HIL simulation model, including the overview of vehicle model, vehicle
ECU network, vehicle plant model, sensors and actuators.

Chapter 5: HIL simulation results and discussions.

Chapter 6: conclusions and future work.

15
2. LITERATURE REVIEW

Type of powertrain architecture plays significant role in improving economy of a HEV.

Series architecture helps in city driving conditions or a drive cycle with a lot of starts
and stops. A parallel vehicle helps in highway drive cycle to improve fuel economy. A
power-split architecture is a combination of parallel and series vehicles but adds the
complexity to vehicle in design and assembly. The power-split architecture also
allows more freedom in control strategy development.

2.1. Power-split Hybrid Electric Vehicle Configurations

A Power-split HEV can have various configurations of itself. A basic power-split is

with one planetary gear set, which is most commonly used and has being
implemented in Toyota Prius and its various generations [4]. A comparison between
two-mode hybrid transmission and IInd generation of Toyota Hybrid System (THS)
was done by Aimin and Hongwei [5]. The two mode transmission has three planetary
gear sets along with two electric motors and an engine. It was awarded technology
of the year of 2007. The system is complex masterpiece of engineering with four
clutches and four fixed gear modes. It allows good speed range with higher
transmission efficiencies and also gives a choice to reduce the electric motor power
rating at design stage compared to a rigid THS single mode system [5]. Chevrolet
Volt is another commercially available single planetary power-split vehicle. It has
three clutches incorporated as output split to get higher efficiencies compared to
Toyota hybrid system with no clutches. These three clutches allow the vehicle to run
in four different modes. The comparison study of THS single mode and Chevrolet Volt
powertrain was done by graduate students at University of Michigan. The study
found that adding one clutch to Toyota hybrid system improves efficiency of the
vehicle by 16.13% while reducing two clutches from the Volt decreases efficiency by
0.8% [6]. A single planetary gear set can have numerous configurations of power-
split. With addition of each planetary gear set, the degree of freedom increases and
the number of combinations to get desired vehicle velocity and torque also increase.
Dr. Jinming Liu worked on the algorithm to design and identify various configurations
of planetary gear set models and their dynamic equations [7].

16
2.2. HEV Control Strategies

The work on the various control algorithms is one of the most influencing parts of the
HEV’s. The control algorithm affects the power plants performance and eventually
vehicle performance. There is significant amount of research being conducted and a
number of control strategies are developed. Poowanart Poramapojana, a graduate
student from Michigan Tech, has conducted a literature review of control strategies in
his thesis research [8]. A lot of work by various researchers is on stochastic control
and dynamic programming. Dr. Jinming Liu worked on Stochastic Dynamic
Programming (SDP) and Equivalent Consumption Minimization Strategy (ECMS) [9]
and found that these control strategies can improve the fuel economy of a power-
split hybrid vehicle. The limitation of the SDP approach is its amount of computation
required. Though a dynamic programming algorithm is faster compared to standard
stochastic control, long computing time and heavy computational need make them
not suitable for the real-time and real world applications with current generation of
embedded systems. ECMS is more suitable for real world application but it is not
robust enough to apply to a vehicle with no predetermined drive cycle. The Adaptive-
ECMS (A-ECMS) control strategy [10] is very similar to ECMS implemented by
Musardo. He tried to eradicate the problem of less robustness by implementing an
adaptive algorithm. This algorithm collects and calculates the data using GPS before
driving condition is reached. This allows a vehicle to decide control parameters
beforehand. But not all of vehicles are equipped with GPS and the signal distortion in
remote areas can hamper a vehicle’s performance in reducing its fuel economy
rapidly. The A-ECMS algorithm is designed and implemented for a parallel vehicle.
This algorithm gives fuel economy values very close to dynamic programming with
very less computational load. Pontryagin’s Minimum Principle (PMP) is another
control strategy which can be implemented for real-time applications [11]. The PMP
algorithm, when implemented on a power-split vehicle, gives results very close to
the dynamic programming.

17
3. DYNAMIC MODEL OF POWER-SPLIT HEV POWERTRAIN

The power-split HEV powertrain architecture has been studied extensively in

literature. There are several variants of power-split powertrain in market. A single
mode in THS I and II is the simplest system with one planetary gear set [4]. THS III,
also called as Hybrid Synergy Drive, uses 2 planetary gear sets. THS III has a
second carrier gear which is fixed to transmission cover and a compound ring gear
[12]. All the THS systems have a motor (MG2), an engine and a generator (MG1).

3.1. SINGLE PLANETARY GEAR SET

Figure 3.1 shows a single planetary gear set.

Ring Gear

Carrier

Pinion Gear

Sun Gear

Figure 3.1: Single planetary gear set

The Planetary Gear (PG) set was modeled with basic equation of kinematic
constraints given by Willis [13] as shown in Eq. 1.

߱௥ . ܴ௥ + ߱௦ . ܴ௦ = (ܴ௥ + ܴ௦ ). ߱௖ Eq. 1

where ߱௥ , ߱௦ and ߱௖ is angular velocity of ring, sun and carrier, respectively, and ܴ௥
and ܴ௦ are radius of ring and sun gears. This equation defines the overall kinematic

18
behavior of the planetary gear set. The equation allows only 2 speed inputs while the
third one has to follow as per the equation.

The simple planetary gear set, used in THS 2004 system, has been studied by Liu J.
[7] and Poramapojana, P. [8] in their dissertation and thesis, respectively. In this
report the 3rd generation of THS system (THS III), Hybrid Synergy Drive, has been
studied.

3.2. HYBRID SYNERGY DRIVE

Figure 3.2 shows the 2 dimensional diagram of the Hybrid Synergy Drive.

Planetary Gear Reduction

Unit Gear

Sun Gear 1 Sun Gear 2

MG1 MG2

Carrier

Compound
Ring Gear

Figure 3.2: Hybrid synergy drive system

Figure 3.3 shows the THS III system, it has a compound ring gear with 3 sets of
teeth as shown in Figure 3.4. The external gear teeth are used to drive the wheels
using a differential gear set. Figures 3.1, 3.3-3.5 were generated from the CAD
software CATIA V5 to help understanding the complexity of hybrid synergy drive
system. The set on the side of orange planetary gears is that of MG2 and is a
reduction gear unit. The other side is for MG1 having planetary gear unit. One

19
planetary gear set, shown in Figure 3.5 is used as power-split device. The power-
split device has engine connected to carrier, the MG1 is connected to sun gear 1, as
shown in Figure 3.2. This power-split device splits the power provided by the engine
towards the sun gear that has MG1 attached to it and the other side of the
compound ring gear connects to a MG2. The ring gear structure allows 20% size
reduction of power-split device from earlier generation of gear drive [12]. The
compact ring gear removes the chain drive presented in the 2nd generation of THS
system. The other planetary gear set is a MG2 speed reduction device. The carrier of
the planetary gear contains spline which is fixed to the transmission casing.

Figure 3.3: 3D model of synergy drive system

20
 Figure 3.4: 3D model of compound ring gear Figure 3.5: 3D model of THS III planetary
gear set

Table 1 shows that all the gear teeth of the synergy drive system used in Toyota
Prius 2010. The fixed reduction gear allows ring and MG2 to rotate in different
direction with respect to each other. This also reduces number of axis required as the
MG2 will have the same axis as MG1 and the engine [12]. The THS II power-split
drive study done by Jerome Meisel [4] explains that the maximum speed of MG1 in
both direction of rotation is the key factor in deciding the engine start and its
maximum speed. As THS II does not have any speed reduction for MG2, it was
directly connected to ring gear. Hence, the RPM of MG1 would reach the speed limit
faster than needed, as the MG1 experiences higher speeds. Gear reduction of second
planetary gear in hybrid synergy drive system allows ring rotations to be lower and
MG2 rotations to be close to MG1 rotations. The reduction device has gear ratio 2.6
ே௢. ௢௙ ௥௜௡௚ ௚௘௔௥ ଶ ௧௘௘௧௛ ଻଼
ቀ ே௢. ௢௙ ௦௨௡ ௚௘௔௥ ଶ ௧௘௘௧௛
= ଷ଴ = 2.6ቁ as per the number of gear teeth/radius of each

gear. The reduced ring RPM removes the kinematic constrain of MG1 and allows the
vehicle to be propelled in only EV mode at higher speeds. The final gear ratio of THS
III has been reduced compared to THS II from 4.113 to 3.2
ே௢. ௢௙ ௙௜௡௔௟ ௗ௥௜௩௘ ௚௘௔௥ ௧௘௘௧௛ ଻଻
ቀ ே௢. ௢௙ௗ௥௜௩௘௡ ௚௘௔௥ ௧௘௘௧௛
= ଶସ = 3.2ቁ [14]. This reduced final gear ratio allows

lesser power loss occurring due to power conversion from MG1 speed reversing on
highway drive cycles compared to THS II [15].

21
 Table 1: Number of teeth on drive gears

Section of Power- Gear No. of gear

split teeth
Ring 1 58
Power-split side Pinion 1 23
Sun 1 22
Ring 2 78
Reduction side Pinion 2 18
Sun 2 30
Counter Drive and Driven gear 54-55
Other gears
Final Drive and Driven Gear 24-77

22
3.3. Dynamics of THS III Gear Train (Hybrid Synergy Drive)

The assumptions for mathematical modeling of compound planetary gear set are as
follows. The clock-wise rotation of ring, carrier and sun gear is considered as positive
while the counter clockwise is considered as negative. The inertia of ring, sun and
carrier arm is considered while pinion inertia has been neglected as it is a very small
value. The pinion experiences force which is transmits towards ring/sun depending
upon the direction of energy flow. The THS III has 2 sections, one is generator side
which is a complete single planetary gear set with all subscript 1 and another is
reduction gear set with subscript 2. The equations for Free Body Diagram (FBD)
shown in Figure 3.6, Figure 3.7 and Figure 3.8 are given in Eq. 2-Eq. 5

‫ܫ‬௦ଶ ߱ሶ ௦ଶ = ‫ܨ‬ଶ . ܴ௦ଶ െ ܶ௦ଶ Eq. 2

‫ܫ‬௥ ߱ሶ ௥ = ‫ܨ‬ଵ . ܴ௥ଵ + ‫ܨ‬ଶ . ܴ௥ଶ െ ܶ௥ Eq. 3

‫ܫ‬௖ଵ ߱ሶ ௖ଵ = ܶ௖ଵ െ ‫ܨ‬ଵ . ܴ௥ଵ െ ‫ܨ‬ଵ . ܴ௦ଵ Eq. 4

‫ܫ‬௦ଵ ߱ሶ ௦ଵ = ‫ܨ‬ଵ . ܴ௦ଵ െ ܶ௦ଵ Eq. 5

The I୰ , Iୡଵ , Iୱଵ and Iୱଶ are mass moment of inertia of ring, carrier 1, sun gear 1 and
sun gear 2 with units of ݇݃ ή ݉ଶ . ɘሶ୰ , ɘሶୡଵ , ɘሶୱଵ and ɘሶୱଶ are the angular acceleration of
ring, carrier 1, sun gear 1 and sun gear 2 with units of ‫݀ܽݎ‬/‫ ݏ‬ଶ . T୰ , Tୡଵ , Tୱଵ and Tୱଶ are
torque of ring, carrier 1, sun gear 1 and sun gear 2 with units of Nm. Fଵ is the
internal force acting on ring and sun gear 1 via pinion gear on generator side and Fଶ
is internal force acting on ring and sun gear 2 due to pinion gear on motors side of
the compound gear set. There is no equation for the carrier 2 as it is fixed to ground.

Figure 3.6: Free body diagram of sun gear 1 (S1) and sun gear 2 (S2)

23
Figure 3.7: Free body diagram of compound ring gear

Figure 3.8: Free body diagram of carrier gear

24
3.4. Modified Hybrid Synergy Drive

Figure 3.9 shows the modified hybrid synergy drive system. A simple clutch has been
added to carrier 1. This allows the carrier on planetary gear side to be grounded at
will and adds an extra dimension in control functionality of the vehicle as MG1 and
MG2 both act as motors.

Planetary Gear
Reduction Gear
Unit
Unit

Sun Gear 2
Sun Gear 1

MG
MG2
1

Carrier

Clutch
Compound
Ring Gear

Figure 3.9: Modified hybrid synergy drive

The study done by Xiaowu and his colleagues [6] showed that an added clutch to the
THS 2004 system helps in improving the economy by 16.132% compared to simple
planetary gear train. Figure 3.9 has added a clutch which allows the vehicle to
operate in 2 modes, Electric Only (EV) mode and Hybrid mode. As the carrier gear 1
gets fixed, it provides the reaction torque necessary so that power-split doesn’t occur
and planetary gear acts as reduction device. Hence, the route of the MG1 power flow
is restricted to flow towards ring and then towards the wheels. The EV only mode
would allow vehicle to be moved without any engine operation. The reduction gear
set removes kinematic constrain (Eq.1) imposed by the MG1 speed as long as the

25
clutch is engaged. The gear ratios show that the gear reduction unit has a gear ratio
ேೝమ ଻଼ ேೝభ ହ଼
of 2.6 ቀ = ቁ almost equaling to planetary gear ratio of 2.6363 ቀ = ቁ.
ேೞమ ଷ଴ ேೞభ ଶଶ

3.5. Dynamic Model of Hybrid Synergy Drive with Clutch

Two planetary gear sets can be reduced to 1 planetary gear set as carrier 2 is
grounded. Now if carrier 1 is grounded as well, the ߱௖ଵ =0 and the sun gear 1 can be
related to the ring gear by Eq. 6

ோೞభ ேೞభ
߱௥ = െ߱௦ଵ . = െ߱௦ଵ . = െ߱‫ݏ‬1. ߩ1 Eq. 6
ோೝభ ேೝభ

where, ܰ௦ଵ and ܴ௦ଵ is the number of teeth and radius of sun gear 1. ܰ௥ଵ and ܴ௥ଵ are
the number of teeth and radius of compound ring gear on MG1 side. As sun gear is
related via pinion and internal gear of ring, the direction of rotation of both the sun
gears and compound ring gear are opposite. For all dynamics, ߱௦ଵ = ߱௚ where ߱௚ is
the angular velocity of the generator and ߱௦ଶ = ߱௠ is the angular velocity of the
motor. No elastic losses between motor/generator shafts and gears are considered.

3.5.1. Electric Only mode (EV)

In EV mode, the clutch of carrier gear 1 is fixed to the ground creating the reaction
torque when MG1 acts as a motor. The planetary gear turns to 2 sets of reduction
gears as shown in Figure 3.10.

26
 Figure 3.10: Power-split powertrain in electric only mode

3.5.1.1. MG2 Dynamics in EV mode

The free body diagram of MG2 is shown in Figure 3.11.

ܶ௦ଶ

‫ܫ‬௠

߱௠ ܶ௠

Figure 3.11: Free body diagram of MG2

‫ܫ‬௠ ߱ሶ ௠ + ܶ௠ = ܶ௦ଶ Eq. 7

Substituting ܶ௦ଶ in Eq. 2 to get the value of internal pinion force ‫ܨ‬ଶ
27
 (ூೞమ ାூ೘ )ఠሶೞమ ା ்೘
‫ܨ‬ଶ = Eq. 8
ோೞమ

3.5.1.2. MG1 Dynamics in EV mode

The free body diagram of MG1 is shown in Figure 3.12.

ܶ‫ݏ‬1

‫݃ܫ‬

߱݃ ܶ݃

Figure 3.12: Free body diagram of MG1

‫ܫ‬௚ ߱ሶ ௚ + ܶ௚ = ܶ௦ଵ Eq. 9

Substituting ܶ௦ଵ in eq. 5 to get the value of internal pinion force

൫ூೞభ ାூ೒ ൯ఠሶೞభ ା ்೒

‫ܨ‬ଵ = Eq. 10
ோೞభ

Substituting values ‫ܨ‬ଵ and ‫ܨ‬ଶ into eq. 3, where ߱ሶ ௦ଵ = ߱ሶ ௚ considering no elastic losses
ோೞభ
in carrier gear. The ratio of sun and ring gear at MG1 and MG2 is given by = ߩଵ ,
ோೝభ
ோೞమ
= ߩଶ . As both are reduction gear sets, ring gear is related to motor and generator
ோೝమ

angular velocities by ߱ሶ ௥ = െ߱ሶ ௠ ߩଶ = െ߱ሶ ௚ ߩଵ . The torque acting on ring gear due to
்ೢ
torque at wheels is given by ܶ௥ = . This is the reaction torque due to the resistance
ீ೏

offered by vehicle dynamics at ring gear. The EV mode is given by Eq.11,

௠௥ మ ൫ூೞభ ାூ೒ ൯ (ூೞమ ାூ೘ ) ்೒ ் ்

ቄீ మ + ‫ܫ‬௥ െ మ
െ మ
ቅ ߱ሶ ௥ = ఘ + ఘ೘ െ ீೢ Eq. 11
೏ ఘభ ఘమ భ మ ೏

where ݉ is the mass of vehicle in kg, ‫ ݎ‬is the radius of the wheels in meters, ‫ܩ‬ௗ is
௠௥ మ
the gear ratio between the wheels and ring gear. The term is vehicle inertia
ீ೏ మ

acting on the ring gear. ‫ܫ‬௥ is ring inertia. As the angular velocities of ring, MG2 and

28
MG1 are related by (Eq.6), a single equation is sufficient to get the wheel velocity,
and source angular velocities.

3.5.2. Hybrid Mode

As the clutch of carrier is disengaged from ground, the carrier can experience force
exerted by both motors. This exerted force also helps in starting the engine and
driving it to idling speed. As soon as engine reaches idling speed and switches to
power mode, vehicle switches to hybrid mode. In hybrid mode, all the three units of
planetary gear set are connected. The dynamics of MG2 remains same while that of
MG1 would change. Since engine is connected to carrier, the dynamics of the engine
is added to the group of equations developed for EV mode. Figure 3.13 shows the
power split device in hybrid mode. Green arrows indicate that power is given to the
vehicle and red arrows indicate that electrical power is generated by MG1.

Figure 3.13: Power-split powertrain in hybrid mode

3.5.2.1. Engine Dynamics in Hybrid Mode

Assume that engine provides a torque in clockwise direction. The torques
experienced by engine is shown in Figure 3.14 with free body diagram at engine
shaft. The dynamics of engine can be given by Eq. 12.

29
 Figure 3.14: Free body Diagram of engine connected to carrier

(‫ܫ‬௘ + ‫ܫ‬௖ଵ )߱ሶ ௘ = ܶ௘ െ ‫ܨ‬ଵ (ܴ௥ଵ + ܴ௦ଵ ) Eq.12

3.5.2.2. MG1 Dynamics in Hybrid Mode

The dynamics of MG1 changes as it is acting as a generator absorbing part of engine
torque. The free body diagram of MG1 shown in the Figure 3.15 is used to form the
dynamic equation

‫ܫ‬௚ ߱ሶ ௚ = ܶ௦ଵ + ܶ௚ Eq. 13

‫ܫ‬௚ ߱ሶ ௚ = ‫ܨ‬ଵ . ܴ௦ଵ + ܶ௚ Eq. 14

߱௚
‫ܫ‬௚

ܶ௚ ܶ௦ଵ

Figure 3.15: Free body diagram of MG1 in hybrid mode

Eq. 1 mentioned before is for a simple planetary gear set. The equation can be
modified to get relation between the angular velocities of MG1, MG2 and engine for
THS III. The new equation is obtained by substituting ߱௥ = െ߱௠ ߩଶ in Eq. 1 as shown
in Eq. 15.

30
 ଵାఘభ ఘ
߱௚ = ቀ ቁ . ߱௘ + ߱௠ ቀఘమ ቁ Eq. 15
ఘభ భ

Differentiating Eq. 15 and substituting into Eq. 14 gives the new value of ‫ܨ‬ଵ for
hybrid mode.

భశഐభ ഐ
൫ூೞభ ାூ೒ ൯ቀ ቁఠሶ೐ ାఠሶ೘ ቀ మ ቁି ்೒
ഐభ ഐభ
‫ܨ‬ଵ = Eq. 16
ோೞభ

Substituting ‫ܨ‬ଵ value into Eq. 12 gives Eq. 17.

ଵାఘభ ଶ ଵାఘభ ఘ ଵାఘభ

൜(‫ܫ‬௘ + ‫ܫ‬௖ ) + ൫‫ܫ‬௦ଵ + ‫ܫ‬௚ ൯ ቀ ቁ ൠ ߱௘ሶ + ቄ൫‫ܫ‬௦ଵ + ‫ܫ‬௚ ൯ ቀ ቁቅ ߱ሶ ௠ ቀఘమ ቁ = ܶ௘ + ܶ௚ ቀ ቁ
ఘభ ఘభ భ ఘభ

Eq.17

This gives one equation in terms of engine angular acceleration and MG2 angular
acceleration. To solve 2 unknowns, the value of ‫ܨ‬ଵ and ‫ܨ‬ଶ are substituted into ring
dynamic equation in Eq. 3. ‫ܨ‬ଶ value will be similar to the value calculated in EV mode
as the carrier 2 is still grounded and MG2 is acting as motor. After substitution and
்ೢ
let ܶ௥ = , ߱ሶ ௥ = െ߱ሶ ௠ ߩଶ , Eq. 3 becomes Eq.18.
ீ೏

ଵାఘభ ௠௥ మ ൫ூೞభ ାூ೒ ൯ (ூೞమ ାூ೘ ) ்೒ ் ்

െ ቄ൫‫ܫ‬௦ଵ + ‫ܫ‬௚ ൯ ቀ ቁቅ ߱௘ሶ െ ቄ ீ మ + ‫ܫ‬௥ െ െ ቅ ߱ሶ ௠ ߩଶ = െ ఘ + ఘ೘ െ ீೢ
ఘభ ೏ ఘభ మ ఘమ మ భ మ ೏

Eq.18

Based on Eq.17 and Eq.18, the state space model in matrix form can be created as
shown in Eq. 19.

ଵାఘ ଶ ଵାఘ ఘ
൜(‫ܫ‬௘ + ‫ܫ‬௖ ) + ൫‫ܫ‬௦ଵ + ‫ܫ‬௚ ൯ ቀ ఘ భ ቁ ൠ ቄ൫‫ܫ‬௦ଵ + ‫ܫ‬௚ ൯ ቀ ఘ భ ቁቅ ቀఘమ ቁ ߱ሶ
൦ భ భ భ
൪൤ ௘ ൨ =
ଵାఘభ ௠௥ మ ൫ூೞభ ାூ೒ ൯ (ூೞమ ାூ೘ ) ߱௠ሶ
െ ቄ൫‫ܫ‬௦ଵ + ‫ܫ‬௚ ൯ ቀ ఘ ቁቅ െ ቄ ீ మ + ‫ܫ‬௥ െ మ
െ మ
ቅ ߩଶ
భ ೏ ఘభ ఘమ

ଵାఘభ ܶ௘
1 ቀ ቁ 0 0 ܶ௚
ఘభ
቎ ଵ ଵ ଵ ቏ ൦ܶ ൪ Eq. 19
0 െఘ ఘమ
െீ ௠
భ ೏ ܶ௪

The equation Eq.19 is of ߙܺሶ = ߚܷ form which can be written as ܺሶ = ߙ ିଵ ߚܷ. The
complete state equation is written as Eq.20.
31
 ߱ሶ 0 0 ߱௘
൤ ௘ ൨=ቂ ቃቂ ቃ
߱௠ሶ 0 0 ߱௠
ିଵ
1 + ߩଵ ଶ 1 + ߩଵ ߩଶ
‫ۍ‬ቊ(‫ܫ‬௘ + ‫ܫ‬௖ ) + ൫‫ܫ‬௦ଵ + ‫ܫ‬௚ ൯ ൬ ൰ ቋ ൜൫‫ܫ‬௦ଵ + ‫ܫ‬௚ ൯ ൬ ൰ൠ ൬ ൰ ‫ې‬
‫ێ‬ ߩଵ ߩଵ ߩଵ ‫ۑ‬
+‫ێ‬ ‫ۑ‬ ‫כ‬
1 + ߩଵ ݉‫ ݎ‬ଶ ൫‫ܫ‬௦ଵ + ‫ܫ‬௚ ൯ (‫ܫ‬௦ଶ + ‫ܫ‬௠ )
‫ێ‬ െ ൜൫‫ܫ‬௦ଵ + ‫ܫ‬௚ ൯ ൬ ൰ൠ െ ቊ ଶ + ‫ܫ‬௥ െ െ ቋ ߩଶ
‫ۑ‬
‫ۏ‬ ߩଵ ‫ܩ‬ௗ ߩଵ ଶ ߩଶ ଶ ‫ے‬

ଵାఘభ ܶ௘
1 ቀ ቁ 0 0 ܶ௚
ఘభ
቎ ଵ ଵ ଵ ቏ ൦ܶ ൪ Eq. 20
0 െ െ ௠
ఘభ ఘమ ீ೏
ܶ௪

If the output vector is selected the same as state vector, the output equation for the
state space model would be Eq.21.

ܶ௘
߱௘ 1 0 ߱௘ 0 0 0 0 ܶ௚
ቂ߱ ቃ = ቂ ቃቂ ቃ + ቂ ቃ൦ ൪ Eq. 21
௠ 0 1 ߱௠ 0 0 0 0 ܶ௠
ܶ௪

The output of ߱௘ and ߱௠ in hybrid mode can be found by solving this state equation.
The angular velocity of wheels ߱௪௛௘௘௟ , can be found using reduction ratio ‫ܩ‬ௗ . While
the angular velocity of the generator, ߱௚ , can be found using Eq. 15.

32
4. HIL SIMULATION VEHICLE MODEL

4.1. Overview of Vehicle Model

The vehicle plant model for the power-split HEV was built based on a parallel vehicle
model provide by dSPACE Inc. The vehicle plant model can be classified into four
categories.

SoftECUs: contains all the ECUs in soft model and all the decision making algorithms.

The Plant: has all the vehicle and subsystem level models to act precisely as vehicle
and its subsystems.

The Sensors and Actuators: has all real time interface blocks needed to simulate and
detect incoming and outgoing signals.

The Environment: simulate the driver and its response, road condition, gear
selections Accelerator Pedal Position (APP) and Brake Pedal Position (BPP).

Figure 4.1 gives the overview of the vehicle model developed in Simulink
environment.

33
 Engine

Battery

Engine Torque
Ambient Temperature and Pressure

MG1 Current MG2 Current

Engine Speed

MG1 Generator MG2 Motor

Vo
Volta
oltage
olt
Voltage
MG1 Torque MG2 Speed
MG2
MG1 Speed
peed Torque
Powersplit
Differential Tire
Speed Torque
Vehicle
Dynamics
Road
Vehicle Speed
Slope
Environment Driver

Figure 4.1: Overview of the vehicle model

34
4.2. Vehicle ECU Network

ECU is the brain of the vehicle which is involved in every decision making process.
The vehicle ECU network is structured to two levels of control. The higher level is
Hybrid Vehicle Controller (HVC). The lower level is gasoline engine controller and
motor controllers. The HVC level is the prime decision making level. HVC subsystem
is divided into “Hybrid mode on/off,” “Brake control,” “Torque request coordination,”
“ICE start/stop,” “Clutch control” and “Drive management.”

“Hybrid mode on/off” decides whether the vehicle is to be driven in EV mode or HEV
mode. “Brake control” decides the amount of brake energy to be regenerated using
MG2 while the rest being passed on to friction brakes. The “Torque request
coordination” is calibrated as per APP and maximum torque curve of MG2 and
engine. The “Clutch control” is a simple model which works with “Drive
management” and vehicle mode selection. The clutch is used to ground engine
carrier arm. The clutch is engaged when vehicle is to be run in EV mode. “Drive
management” is a crucial subsystem which decides the vehicle mode selection,
engine mode selection and the distribution of the torque based on these modes. The
vehicle plant model is designed to be EV/HEV. The vehicle has two modes with
planetary gear set at its core. The first mode is the Electric only (EV) mode and the
second mode is power-split HEV mode. ICE has 3 modes, (1) off (2) idle and (3)
power mode. All these modes are controlled by a Stateflow control algorithm.

Figure 4.2 shows the control algorithm for mode selection process. The model was
divided into four blocks. Vehicle starts with engine off (ICE mode=0) and electric
only mode on (vehicle mode =0). If the torque request is greater than 0 as well as
the State Of Charge (SOC) is lower than SOC lower limit (60%), mode selection
switches to ICE mode=1. This is ICE idling mode. The clutch is released to start the
engine. Vehicle goes to hybrid mode=1 only when vehicle speed is greater than
20mph. This is because the engine operates in very high BSFC region at low rpm,
which gives overall low efficiency. Engine can also be in power mode if SOC falls
below SOC_emg along with battery temperature conditions for safety reasons.
SOC_emg is very low SOC region and battery needs to be charged immediately. If

35
battery temperatuUH IDOO RXW RI VDIH RSHUDWLQJ UHJLRQ Û&EDWWHU\
WHPSHUDWXUHÛ& , the engine provides traction power.

If the torque request falls below 0, vehicle is braking. The engine changes to idling
mode and the vehicle changes to electric mode. If braking time is longer than 1
second, the clutch is engaged, engine is turned off, and MG1 along with MG2 provide
regenerative braking. If engine charges battery to the upper limit and battery
temperature is within operating limits, engine stops and both electric motors provide
required torque. Vehicle operates in electric only mode.

Input state

[((Trq_request_vehicle>0)&&(SOC<=SOC_lowlimit))
EngineOff
||(T_Bat>=T_Bat_uplimit)||(T_Bat<=T_Bat_lowlimit)]
du:ICE_Mode=0;
du:Vehicle_Mode=0;
du:engine_rpm_state=0;
PreFreqStart
en:t_PreFreqStart=Time;
[((Time-t_PreFreqStop)>=1)] du:ICE_Mode=0;
du:Vehicle_Mode=0;
EngineIdle
du:ICE_Mode=1;
du:Vehicle_Mode=0;
[(Time-t_PreFreqStart)>=1]
en:t_PreFreqStop=Time;

[(Trq_request_vehicle<=0)||((SOC>=SOC_uplimit)
&&(T_Bat<T_Bat_uplimit)&&(T_Bat>T_Bat_lowlimit))]

[((Trq_request_vehicle>0)&&(vel_kmph>20))
||(SOC<=SOC_emg)||(T_Bat>=T_Bat_uplimit)
||(T_Bat<=T_Bat_lowlimit)]

EnginePower
du:ICE_Mode=2;
du:Vehicle_Mode=1

Figure 4.2: Control strategy to select vehicle mode

36
The torque request distribution was done in both modes. In electric mode, as clutch
is engaged and carrier 1 is locked, MG1 along with MG2 can produce torque to propel
the vehicle. The torque request was distributed to ratio of MG2:MG1=1:1 as both the
motors are of equal rating and the gear ratio are equal. While starting the engine,
MG1 is starting motor as well. MG1 has to produce torque in clockwise direction. As
soon as the ICE mode changes to idling, the MG2 has to provide the entire torque
request. MG1 compensates the engine speed with respect to vehicle speed so that
engine idling speed of 800 rpm is maintained.

In hybrid mode, torque split is decided on rule-based control algorithm. The RPM of
engine decides the torque of engine and is limited by the maximum MG1 RPM as well
as current vehicle speed as per equation Eq.15. The purpose of hybrid mode is to
recharge battery while operating engine in optimum range to get maximum fuel
efficiency for given condition. For hybrid mode torque distribution, power request
from drivers is split into engine power request and MG2 motor power request. The
split ratio is determined based on current SOC of battery using a lookup table.
Engine provides most power at low SOC while charge sustaining. Once the power is
split, engine speed is decided from a look-up table containing optimum power to
speed curve. The speed from loop-up table divides power request for engine to give
engine torque request. The e-motor rpm coming from MG2 motor divides its power
request to give motor torque request. In hybrid mode, engine operates in two states,
battery charge sustaining and battery charging. In charge sustaining state, engine
provides torque to the wheels completely via electrical and mechanical path. In
mechanical path the torque is provided through the gears of power-split device and
reduction device. In electrical path, torque goes through MG1 generated and is
provided to MG2 for propulsion. Thus, very low or no power is provided by the
battery in charge sustaining mode. In charging mode, the engine provides excess
power so that the torque requirement for mechanical and electrical path is fulfilled
along with battery SOC build-up. In summary, the power of the battery is controlled
in three states: charge depletion state in electric only mode, charge sustaining and
charging state in hybrid mode of vehicle.

37
The Engine Control Unit (ECU) and MG1/MG2 controllers are at second level. Engine
control unit used in this case of HIL simulation was a SoftECU. The engine SoftECU is
provided by dSPACE. The ECU is presented in ASM library of dSPACE. It is used to
decide the injection pressure, mass of fuel, mass of induced air, turbocharger
pressure, EGR control, throttle position and various other parameters for engine
plant model. DS2211 board is used to simulate the sensor and actuator signals. The
various blocks needed for the interface can be found in dSPACE library.

The MG1 controller is a softECU used to control MG1 while MG2 controller is a
hardware ECU built in MicroAutoBox. MG1 and MG2 controllers are used to control
their respective motors on the basis of torque request calculated by hybrid ECU in
electric only mode. In hybrid mode, MG2 is controlled with torque request as it is the
primary motor while MG1 is controlled by engine RPM request as its primary function
is to compensate the engine RPM with respect to vehicle speed.

38
4.3. Vehicle Plant Model

The plant model of the vehicle has a gasoline engine, 2 Permanent Magnet
Synchronous Motors (PMSM), a power-split drive train, a battery model, vehicle
dynamics model and environment subsystem. Figure 4.3 shows the architecture of
the vehicle.

Power
control
unit

Planetary Battery
Gear Unit

Engine

MG
MG2
1

Reduction
Gear

Figure 4.3: Vehicle architecture

4.3.2. Engine Model

A 2.9 liter, 6 cylinders engine is modeled in Simulink platform. The model is divided
into 5 subsystems as shown in Figure 4.4. “Fuel system” controls all the fuel pump,
fuel metering, and injection pressure. “Air path” has intake and exhaust manifold
along with EGR and turbocharger. The “Piston engine” has the cylinder inlet, torque
generated due to combustion. Finally, the “Coolant subsystem” and “Exhaust
subsystem” deal with engine cooling and exhaust emissions. The engine model,

39
designed by dSPACE, is equipped with turbocharger and EGR. For the power-split
vehicle, the engine was equipped with just EGR. Turbocharger was kept off.

Injection Timing, Exhaust

Duration, Crank angle Manifold
Fuel System
Mass of Fuel, Exhaust Lambda
Injection Time

Engine
Engine ECU
Combustion

Coolant
Intake
Temperature,
Manifold
Throttle Position, EGR % Intake manifold Mean torque
and Turbocharger Pressure and temp
Engine
Torque

Figure 4.4: dSPACE engine model

4.3.2. Model of Electronic Components

4.3.2.1. Motor Model
The planetary gear set of hybrid synergy drive train has two motors attached to two
sun gears. The MG1 attached to sun gear 1 and the MG 2 is attached to sun gear 2.
The motors selected for this hybrid vehicle are two 43kw Permanent Magnet
Synchronous Motors. Motor is selected on the basis of its high efficiency and high
torque performance. Figure 4.5 shows motor efficiency map. Mathematical model for
motors is given by Eq.23-25 [16]

ௗ௜೏
‫ܮ‬ௗ = ‫ݒ‬ௗ െ ܴ௦ ݅ௗ + ‫ܮ‬ௗ ߱௘ ݅௤ Eq.23
ௗ௧

ௗ௜೜
‫ܮ‬௤ = ‫ݒ‬௤ െ ܴ௦ ݅௤ + ‫ܮ‬௤ ߱௘ ݅ௗ െ ߰௣ ߱௘ Eq.24
ௗ௧

ଷ
ܶ‫߰[݌ = ݍݎ‬௣ ݅௤ + ൫‫ܮ‬ௗ െ ‫ܮ‬௤ ൯݅௤ ݅ௗ ] Eq.25
ଶ

where ‫ܮ‬ௗ ܽ݊݀ ‫ܮ‬௤ are inductance along d-axis and q axis of reference. ‫ݒ‬ௗ ܽ݊݀ ‫ݒ‬௤ are
voltage along d-axis and q axis of reference. ݅ௗ ܽ݊݀ ݅௤ are current in d-axis and q axis
of reference. ܴ௦ is stator resistance. ߱௘ is the angular velocity of motor, ߰௣ is the flux
induced by permanent magnet. ‫ ݌‬is the number of poles.

40
 Figure 4.5: MG2 and MG1 motor efficiency map

Figure 4.6 shows the block diagram of electric component model. A PMSM motor
needs three-phase current for its coils to generate the required torque. The three-
phase current is generated by a three-phase inverter. For this project, IGBT switches
are used in the three-phase conversion process of inverter. The Pulse Width
Modulation (PWM) signal coming from motor controller is used to control three-phase
inverter. The duty cycle of the PWM signal decides the amount of current to be sent
to motors. By controlling the duty cycle, the speed of PMSM motor is controlled.

41
 MG2 Controller
MicroAutoBox

Duty Cycle, Rotor Position RPM

Stator
Torque and
Voltage
Load Current
Three Phase a, b, c Electric motor
Inverter MG2 Model MG2

Stator
Voltage
Three Phase Inverter a, b, c Electric motor
MG1 Model MG1
Torque and
Load Current
Rotor Position
Torque,
stator current a,b,c
MG1
Duty Cycle
Controller

Simulator

Figure 4.6: dSPACE electric components model

The MG2 motor is controlled using MicroAutobox in dSPACE HIL environment. The
plant model sends motor RPM, three-phase current to the controller and motor
torque request is send by hybrid ECU. The MG1 plant model is similar to MG2 plant
model in all aspects as the motor specifications are same. The controller for MG1
used is a soft controller. The MG1 controller is very similar to MG2 controller.
However, MG1 controller is implemented with plant model in dSPACE simulator not in
MicroAutoBox.

The vehicle speed control is based on MG2 speed control and the compensatory
speed developed by MG1 for engine speed variation. As vehicle does not have any
automatic or manual transmission, MG1 speed variation is very crucial. This allows
the engine to be used at any speed irrespective of vehicle speed but limited by the

42
MG1’s maximum speed in both directions. Control of vehicle is done by controlling
the MG1 speed in Hybrid mode. The MG2 motor is controlled using current control
method which is proportional to torque generated by motor.

For both motors, the controller was borrowed from the ASM library of dSPACE
software module. Controller used is a simple PI controller to control a motor in 3
operating modes (1) speed control, (2) current control and (3) voltage control. In
current control mode, torque request is converted to q-axis current and is controlled
by a PI controller. While in speed control, the RPM is converted to q-axis current and
then controlled by the PI controller. This model is developed by the dSPACE Inc. The
controller works in Clark and Park transformation, where ‫ܫ‬ௗ and ‫ܫ‬௤ current/voltage
are controlled in d-q axis. The transformation is shown in Eq. 26.

ଶగ ଶగ
‫ܫ‬ௗ cos ߠ cos(ߠ െ ) cos(ߠ + ) ‫ܫ‬௔
ଷ ଷ
ଶ
൥‫ܫ‬௤ ൩ = ൦ sin ߠ sin(ߠ + ) ൪ ൥‫ܫ‬௕ ൩ Eq. 26
ଶగ ଶగ
ଷ sin(ߠ െ )
0 ଷ ଷ ‫ܫ‬௖
0.5 0.5 0.5

The primary role of MG1 is to act as a speeder in hybrid mode of vehicle. The angular
velocity of motor is used to keep the speed of engine at optimum level so as to get
maximum efficiency over a drive cycle. But MG1 also acts as a load motor in EV
mode which is controlled in current control mode. The transition of the current
control to speed control takes place in EV mode of vehicle. Here, the engine can be
in idling and clutch is disengaged. When engine is to be started, MG1 provides the
necessary torque along with MG2 in speed control mode. MG2 provides complete
torque to vehicle to fulfill the torque request. The MG1 controller only works as
current controller when engine is off. The Hybrid ECU decides which mode the
controller must act so as to get maximum efficiency of the vehicle.

43
4.3.2.2 Battery Model
The battery model was borrowed from ASM library of dSPACE Inc. The model
consists of maximum 100 cells with rating of the module as shown in the Table 2.

Table 2: Battery Specification

Battery Specs

Type Li-ion

Number of cells 100

Voltage 360V

Peak Current 250A

The battery management for two motors was added to the battery subsystem. As
both MG2 and MG1 are used as motor and generator, the right connection between
the outputs of MG2 and MG1 can save the SOC for future need of drive cycle. In
Hybrid mode, the MG1 acts as a generator while the MG2 is a motor drawing the
current from battery. The motor current is three-phase so the load current is
calculated by summing three phases. The three-phase current is provided by a
three-phase inverter and is directed towards battery to calculate battery current. The
load current of motor MG2 varies with torque requested. The forward modeling of
battery has voltage as output and motor current as input. The battery current can be
manipulated by simple addition of currents as per the equation Eq. 22. This addition
allows MG1 generated current to be sent to MG2 motor and remaining can be sent to
battery for its sustainability. So, the addition decides the battery current ‫ܫ‬௕௔௧ and the
SOC is affected based on the addition of current.

‫ܫ‬௕௔௧ = ‫ܫ‬ெீଶ + ‫ܫ‬ெீଵ Eq.22

44
4.3.3. Development of Power-split Drivetrain Model
The power-split drive train was newly modeled for this vehicle. A modified THS III
was considered for power-split HEV HIL simulation. 2010 Prius has synergy drive
(THS III) built into it. The modeled drive train has 2 planetary gear set
configurations as mentioned earlier. Three components, i.e. the sun-1 carrier-1 and
sun-2, are powered by MG1, gasoline engine, and MG2, respectively.

The model developed in Simulink was based on the equations developed in section 3.
The dynamics of engine, MG1 and MG2 was included in to the dynamics of power-
split device. Hence, no dynamic models for engine or motors were needed. The
torque signals coming from engine, MG2, MG1 and the vehicle resistance coming
from vehicle dynamics are given as inputs to the EV mode block and hybrid mode
block as shown in Figure 4.7.

Electric Only Mode Ring RPM

MG2 RPM

Ring RPM to
Wheel RPM
MG2/MG1 RPM

MG1 RPM

Torque Engine MG2 RPM

Engine RPM Clutch Sig Mode Switch MG1 RPM

Torque MG2 Wheel RPM
Hybrid Mode

Torque MG1 MG2 RPM to MG1 RPM

MG1 and
MG2 RPM
Differential gear
RPM
Torque_Wheels MG1 RPM

Figure 4.7: Power-split model

45
The equations needed for simulation of electric and hybrid modes are of discrete
forms as the simulator works at sample rate of ܶ௦ = 0.001 seconds. So, the model is
executed for each time step of 0.001 seconds. For electric only mode equation,
Eq.11 has one variable, angular acceleration of ring. This has been solved by division
of torque and inertia as shown in equation Eq.27

೅೒ (ೖ) ೅೘ (ೖ) ೅ೢ (ೖ)

ఠೝ (௞)ିఠೝ (௞ିଵ) ି ି ି
ഐభ ഐమ ಸ೏
்ೞ
= ೘ೝ మ ൫಺ೞభ శ಺೒ ൯ (಺ೞమ శ಺೘ )
Eq. 27
ቊ మ ାூೝ ି ି ቋ
ಸ೏ ഐభ మ ഐమ మ

where k is the discrete time step. Three input variables, torque of MG1 ܶ௚ (݇), torque
of MG2 ܶ௠ (݇) and reaction torque from wheels ܶ௪ (݇) in discrete form, were used to
calculate angular acceleration of ring ߱ሶ ௥ (݇). Torque values from MG2 and MG1 are
negative as the torque is generated in 3rd quadrant of PMSM map. Hence, the signs
of MG2 and MG1 torque in Eq.27 are negative so that the resultant torque of ring is
positive and ߱ሶ ௥ (݇) is positive. For electric only mode, the angular velocities of MG1
and MG2 are related to the RPM of wheel by gear ratios ߩଵ , ߩଶ and ‫ܩ‬ௗ . The values of
߱௚ (݇) , ߱௠ (݇) and ߱௪ (݇) were calculated by Eq.28, Eq.29 and Eq.30.

ିఠೝ (݇)
߱௚ (݇) = Eq. 28
ߩ1

ିఠೝ (݇)
߱௠ (݇) = Eq.29
ߩ2

ఠೝ (݇)
߱௪ (݇) = Eq. 30
ீ೏

For hybrid mode, Eq.17 and Eq.18 have two variables. They are angular acceleration
of engine and angular acceleration of MG2. Eq. 17 and Eq. 18 are in continuous form.
They are converted to discrete form as shown in Eq. 33 and Eq. 34.

1 + ߩଵ ଶ ൫߱௘ (݇) െ ߱௘ (݇ െ 1)൯ 1 + ߩଵ ߩଶ

ቊ(‫ܫ‬௘ + ‫ܫ‬௖ ) + ൫‫ܫ‬௦ଵ + ‫ܫ‬௚ ൯ ൬ ൰ ቋ× + ൜൫‫ܫ‬௦ଵ + ‫ܫ‬௚ ൯ ൬ ൰ൠ × ൬ ൰
ߩଵ ܶ௦ ߩଵ ߩଵ
൫߱௠ (݇) െ ߱௠ (݇ െ 1)൯
×
ܶ௦

46
 ଵାఘభ
= ܶ௘ (݇) + ܶ௚ (݇) × ቀ ቁ Eq. 33
ఘభ

1 + ߩଵ ൫߱௘ (݇) െ ߱௘ (݇ െ 1)൯ ݉‫ ݎ‬ଶ ൫‫ܫ‬௦ଵ + ‫ܫ‬௚ ൯ (‫ܫ‬௦ଶ + ‫ܫ‬௠ )

െ ൜൫‫ܫ‬௦ଵ + ‫ܫ‬௚ ൯ ൬ ൰ൠ × െ ቊ ଶ + ‫ܫ‬௥ + + ቋ × ߩଶ
ߩଵ ܶ௦ ‫ܩ‬ௗ ߩଵ ଶ ߩଶ ଶ
൫߱௠ (݇) െ ߱௠ (݇ െ 1)൯
×
ܶ௦

்೒ (௞) ்೘ (௞) ்ೢ (௞)

= െ െ െ Eq. 34
ఘభ ఘమ ீ೏

For hybrid mode the generator speed ߱௚ (݇) is calculated by Eq. 35,

ଵାఘభ ఘ
߱௚ (݇) = ቀ ቁ . ߱௘ (݇) + ߱௠ (݇) ቀఘమ ቁ Eq. 35
ఘభ భ

whereas wheel RPM is calculated using Eq. 36

ିఠ೘ (݇)‫כ‬ఘమ
߱௪ (݇) = Eq. 36
ீ೏

All the RPM values calculated using Eq. 28 to Eq. 36 were used in the mode selection
block as shown in Figure 4.7. The mode selection is done on the basis of clutch signal
while clutch signal is decided on the basis of ICE mode. If ICE mode is in idling (ICE
mode=1) or power (ICE mode=2), clutch is released from ground and carrier 1 is
free to rotate. The change in vehicle mode was done with a multiport switch with
clutch signal being switching signal. The output RPM values of MG2, MG1 and engine
were sent as feedback to respective plant models. RPM value of wheel was used to
calculate the vehicle velocity.

47
4.3.4. Vehicle Dynamics
Vehicle dynamics model has all the resistive forces acting on wheels of vehicle. The
vehicle experiences rolling resistance, aerodynamics drag, gradient resistance, brake
force and tractive force as shown in Figure 4.8.

Figure 4.8: Vehicle dynamics

The rolling resistance acts due to rolling of the wheels. The rolling wheel has its own
weight acting in downward direction which creates small deformation of the wheel.
Ideally, a wheel should be touching the road surface tangentially with minimum
surface contact. Due to surface contact, these wheels offer resistance to the
movement in rotational direction. The rolling resistance is given by Eq. 37, where ݉
is the mass of vehicle in kg, ݃ is the acceleration due to gravity in m/s2, ߠ is road
gradient in degree, ݂௥ is friction factor.

‫ܨ‬௥௢௟௟ = ݉݃ cos ߠ ݂௥ Eq. 37

Aerodynamic drag is the fluidic resistance experienced by the vehicle when it passes
through a medium. In the case of automobiles, the vehicle passes through the
medium of air with density of 1.2kg/m3. The vehicle experiences a drag in opposite
direction of vehicle movement. This drag can increase or decrease with the
increase/decrease in wind velocity. For the simulation, the air as a medium was
considered steady. The aerodynamic force is a function of cross sectional area of
vehicle and the square of vehicle velocity as shown in Eq. 38, where ߩ is the density

48
of air in kg/m3, ‫ܥ‬ௗ is the coefficient of aerodynamic drag, ‫ ܣ‬is the cross sectional
area in m2, ‫ ݒ‬is the velocity of vehicle in m/s.

ଵ
‫ܨ‬ௗ௥௔௚ = ߩ‫ܥ‬ௗ ‫ ݒܣ‬ଶ Eq. 38
ଶ

Third type of resistance experienced by the vehicle is the gradient resistance. This
resistance is due to the road slope. The vehicle may be assisted or resisted
depending on the direction of vehicle velocity and the slope. The weight of the
vehicle helps in propelling the vehicle when velocity and slope are in same direction.
The gradient resistance is a function of slope angle and vehicle weight given in Eq.
39. For the simulation, the slope was considered 0.

‫ܨ‬௚௥௔ௗ = ݉݃ sin ߠ Eq. 39

The traction force is the force coming from the powertrain of vehicle towards the
wheels and is the only force propelling the vehicle in forward direction. For power-
split hybrid vehicle, there are three power sources, one engine and two electric
motors. The brake force is the necessary force needed to stop the vehicle. A hybrid
vehicle can apply brake force using the regenerative braking and friction brakes. In
EV mode, both motors apply the regenerative brake while in hybrid mode just MG2
applies the regenerative brake. The friction brake is only applied when needed brake
force is greater than the capacity of regenerative brakes.

The linear dynamics of the vehicle can be represented by Eq. 40. The parameter
values used in simulation are given in Table 3.

ଵ
‫ ܽ݉ = ்ܨ‬+ ݉݃ cos ߠ ݂௥ + ߩ‫ܥ‬ௗ ‫ ݒܣ‬ଶ + ݉݃ sin ߠ Eq. 40
ଶ

Table 3: Constants used in simulation

Sr. No. Variable Value

1 Mass of vehicle in kg 1250
2 friction factor 0.01
3 density of air in kg/m3 1.2

49
 4 coefficient of aerodynamic drag 0.35
5 cross sectional area in m2 2
6 Road gradient 0

4.4. Sensors and Actuators

The sensor and actuator interfaces are used for the MG2 motor since the MG2
controller is a hardware controller, MicroAutoBox. The MG2 motor model
communicates with MicroAutobox via DS 5202 board. Analog outputs to the
MicroAutoBox include three-phase current, a MG2 RPM and MG2 torque request. The
resultant torque applied by PMSM motor needs to be perpendicular to mechanical
rotor position to get maximum efficiency of motor. The HIL simulator uses an
Angular Processing Unit (APU) to capture the angle of rotor position.

4.5. Environment

The environment subsystem has all the maneuverability and the driver signals.
Maneuverability sends the signal of vehicle reference velocity. The driver signals are
APP, BPP, and gear lever selection signals. A PI controller with feed forward and
feedback position is used to control the APP and BPP positions.

50
5. HIL SIMULATION AND RESULT DISCUSSION

5.1. dSPACE HIL Simulator

Any design of power-split configuration and its control strategy needs to be tested.
It’s better if the tests are before manufacturing of components. It will allow and
reduce the time frame for optimum design. To get the test results close to actual
results after manufacturing, real world testing is of great importance. Once a
dynamic model of a vehicle is tested in software, the plant model and its control
algorithms of various systems and subsystems can be tested with latest edition of
simulator from dSPACE. The main plant model is built for dSPACE ECOLine simulator
and Motors is controlled with MicroAutobox II for this testing of vehicle. The ECOLine
simulator has a DS1006 quad core board. The simulator has 3 ECUs with various
configuration points given in document [17]. The simulator contains the digital signal
conditioners, signal processors, various sensor and actuator interfaces. DS5202
board is being used for communication between simulator and MicroAutobox. The
MicroAutobox contains two boards, 1511 and 1512 board and DS1552 piggyback
board. The box acts as a motor controller. The 1512 board contains all the ADC and
Hall Effect sensor interfaces but doesn’t have any Controller Area Network (CAN)
interfaces. The 1511 contains CAN interface as well as serial interface for the
communication with other ECUs. Basic functionality of the simulator is to act as a
vehicle plant while connected to the hardware ECU which controls plant. If hardware
ECUs are not present, softECUs can also be implemented in the simulator with
vehicle plant model. Once model is completed and debugged, the binary simulation
code can be easily generated by auto-code generation software and downloaded to
simulation board. If hardware ECUs are presented in the simulation system, the code
for these controllers can be flashed to corresponding controllers so that the
simulation can be run with hardware-in-the-loop

The model calibration is done in ControlDesk software provided by dSPACE Inc. Once
the model is built in the Simulink environment, it is flashed onto simulator board.
ControlDesk uses variables from *.sdf file which is created while building the model.
The ControlDesk allows calibrating variables in real-time using different instruments.

51
The software is extremely helpful for real-time calibration of variables such as gain
values of PI controllers.

5.2. HIL Simulation Setup

HIL simulation setup was done by connecting the ECOLine simulator and
MicroAutobox. The details of connection are listed in the Table 4. The simulation
computer was connected to the HIL simulator directly while the MicroAutobox was
connected using an Ethernet cable. When registering the simulator platform and
MicroAutobox platform, the simulator can be detected without IP address while the
MicroAutobox needs to be inserted with IP of 141.219.24.56. Figure 5.1 shows the
HIL simulation setup.

Electronic Control Unit MicroAutoBox

Micro- Motor ECU
controller

Signal Output
Conditioning Drivers

Electrical Signal
dSPACE HIL Si
Simulator
l t
Interface

Sensors Actuators

Automotive Plant Model

Figure 5.1: HIL simulation set-up

52
 Table 4 : The connections between HIL simulator and MicroAutoBox

Sr. DS5202 EMH

MicroAutobox II pins Signal
No. Simulator pins

37-pin Female Sub-D

1 Hall 1A D3 Rotor position
Connector P1: Pin20
37-pin Female Sub-D
2 Hall 1A E6 Rotor position
Connector P1: Pin21
37-pin Female Sub-D
3 Hall 1B D4 Rotor position
Connector P1: Pin22
37-pin Female Sub-D
4 Hall 1B D5 Rotor position
Connector P1: Pin23
37-pin Female Sub-D
5 Hall Z E4 Rotor position
Connector P1: Pin24
37-pin Female Sub-D
6 Hall Z E5 Rotor position
Connector P1: Pin25
50-pin Female Sub-D
7 ADC Input-pin Three Phase current
Connector P2: Pin41
50-pin Female Sub-D
8 ADC Input-pin Three Phase current
Connector P2: Pin 25
50-pin Female Sub-D
9 ADC Input-pin Three Phase current
Connector P2: Pin9
50-pin Female Sub-D
10 ADC Input-pin Three Phase current
Connector P2: Pin42
50-pin Female Sub-D
11 ADC Input-pin Voltage signal
Connector P2: Pin10
50-pin Female Sub-D
12 ADC Input-pin Voltage signal
Connector P2: Pin43
50-pin Female Sub-D
13 ADC Input-pin Three Phase current
Connector P2: Pin27
50-pin Female Sub-D
14 ADC Input-pin Three Phase current
Connector P2: Pin11
50-pin Female Sub-D
15 ADC Input-pin Torque request signal
Connector P2: Pin28
50-pin Female Sub-D
16 ADC Input-pin Torque request signal
Connector P2: Pin 12
50-pin Female Sub-D Resolver Excitation- analog position sensor
17
Connector P2: Pin5 a4 signal
50-pin Female Sub-D Resolver Excitation- analog position sensor
18
Connector P2: Pin38 b6 signal
50-pin Female Sub-D analog position sensor
19 Resolver Sine a5
Connector P3: Pin18 signal
50-pin Female Sub-D analog position sensor
20 Resolver Sine a6
Connector P3: Pin 2 signal

53
 50-pin Female Sub-D analog position sensor
21 Resolver Cosine c6
Connector P3: Pin35 signal
50-pin Female Sub-D analog position sensor
22 Resolver Cosine b5
Connector P3: Pin19 signal
50-pin Female Sub-D
23 Digital In 7 T5 Incremental Encoder Signal
Connector P3: Pin27
50-pin Female Sub-D
24 Digital In 8 T6 Incremental Encoder Signal
Connector P3: Pin14
37-pin Female Sub-D
25 PWM Input R3 Duty Cycle
Connector P4: Pin20
37-pin Female Sub-D
26 PWM Input R5 Duty Cycle
Connector P4: Pin21
37-pin Female Sub-D
27 PWM Input S5 Duty Cycle
Connector P4: Pin22
37-pin Female Sub-D
28 PWM Input R6 Duty Cycle
Connector P4: Pin23
37-pin Female Sub-D
29 PWM Input R4 Duty Cycle
Connector P4: Pin24
37-pin Female Sub-D
30 PWM Input S4 Duty Cycle
Connector P4: Pin25

5.3. HIL Simulation Results and Discussion

This report is focused on testing of the modeled power-split device and its
performance in a vehicle environment using HIL simulation. The power-split model
was integrated into vehicle model in Simulink environment. The compiled vehicle
model was loaded onto the simulator processor board while the control model of MG2
was loaded onto to the MicroAutobox. Once the models were downloaded, the
simulation was performed with ControlDesk software for real-time calibration. The
simulation was conducted for an Urban Dynamometer Drive Schedule (UDDS) drive
cycle over a period of 1400 seconds.

54
Figure 5.2 shows the comparison between reference UDDS drive cycle and HIL
simulation results of power-split vehicle. The reference drive cycle is built into the
environment level of vehicle model. The drive cycle is followed precisely at most
time. The simulation output exceeds reference vehicle speed only for a short period
of time starting at 480sec due to rule-based control actions of hybrid mode
controller. The torque provided by engine is in excess to torque request to keep
engine running and charge the battery above 60%. The maximum error of vehicle
speed is 1% in EV mode and 4% in hybrid mode except a short period of time
starting at 480sec, within which an error of 10% is observed. The simulation result in
Figure 5.2 gives a good base to conclude that the vehicle model is working properly
with developed power-split device model and rule-based control strategy.

Figure 5.2 : Vehicle Speed tracking in UDDS

55
Figure 5.3 shows the engine torque request in hybrid mode and the actual amount
generated by the engine. Engine output torque follows closely in most cases except
at highly spiked request time instances as shown in magnified figure. In these cases,
engine is not able to generate enough output torque within such a short time period.
One example is when t=205 seconds. In other cases, engine produces and follows
torque precisely as requested. Throughout the drive cycle, engine is in power mode
only for charge sustaining and charging the battery. Engine is turned off for charge
depletion in electric only mode.

Figure 5.3: Comparison of requested ICE torque and generated ICE torque

56
Figure 5.4 shows the SOC of the battery. As per the control strategy, the primary
function of engine power mode is to increase SOC level to 70% limit. The engine
starts to charge battery as soon as SOC falls below 60%. Figure 5.2 shows that the
velocity profiles are almost repetitive after 350 seconds. Hence, the engine charges
the battery at peak velocity profiles while regenerative strategy shuts the engine off.
The SOC drops in electric only mode significantly to provide necessary vehicle
starting torque. After that, the SOC is maintained around 60%.

Figure 5.4: Comparison of battery SOC with SOC upper limit

57
Figure 5.5 shows the RPM of MG2, primary driving motor. The motor rotates in
counter clockwise direction and the wheels are in clockwise direction. As the wheels
moving forward are positive, motor direction has to be negative. The motor RPM is
recorded in simulation for EV and hybrid mode. The MG2 RPM is directly connected to
the wheel RPM by a gear ratio of 2.6 reduction and 2.2 differential. As a result, the
MG2 has the same profile as that of vehicle speed but in opposite direction.

Figure 5.5: The RPM of primary motor, MG2

58
Figure 5.6 shows the RPM of the generator. The generator acts as a motor in EV
mode and as a generator in Hybrid mode. In EV mode, engine can be off or idling.
Once the clutch is released, MG1 acts as a starter motor along with MG2 to start the
engine. As the RPM of engine in starting and idling mode are limited to 800RPM, MG1
has to act quickly in clockwise direction to help ring in starting the engine. That’s
why there are spikes in MG1 towards positive direction. In EV mode and regenerative
braking, MG1 and MG2 gear ratios are similar (2.6 and 2.63). It can be seen that
velocity profile of MG1 and MG2 are same while braking.

Figure 5.6: The RPM of secondary motor MG1

59
Figure 5.7 shows different ICE modes attended in comparison with vehicle velocity.
The ICE mode=0 is engine off, ICE mode=1 is engine idling, ICE mode=2 is engine
power mode. The rule-based control strategy for ICE mode selection is governed by
SOC percent. When the SOC drops below 60%, the engine starts and it either
operates in charge sustaining mode or in charging mode as per the torque request.
As it is seen in Figure 5.4, SOC level is risen up to 62% by the end of 350 seconds.
After 350 seconds, the drive cycle is almost repetitive. Hence, the engine starts at
the start of each peak and stops when regenerative braking is applied.

Figure 5.7: Comparison of ICE mode change with vehicle velocity

60
Figure 5.8 shows the comparison between engine RPM and MG1 RPM. In hybrid
mode, MG1 compensates engine RPM for given vehicle speed. The engine RPM is
limited by the MG1 maximum speed of 8700 rpm. MG1 RPM is calculated by Eq.15 in
hybrid mode. The RPM profile is close to engine profile with deviation of MG2 profile.
The spikes seen are due to sudden change of MG1 speed from counter clockwise to
clockwise to start the engine.

Figure 5.8: Generator RPM and engine RPM

Figure 5.9 shows the comparison of vehicle mode and ICE mode with rotational
speeds of MG1, MG2, and engine. Vehicle has two modes: Electric only (EV) mode
=0 and Hybrid mode =1. Engine has three modes: power mode (ICE mode=2), idling
mode (ICE mode=1), and engine off mode (ICE mode=0). Vehicle mode changes to
Hybrid mode only when the ICE is in power mode (ICE mode=2). The powertrain
components follow the primary kinematic equation given in Eq. 15 in hybrid mode.
The powertrain components connected directly to wheels via gear ratios. In electric
only mode, motors follow vehicle velocity profile while engine rpm is 0. The
simulation result shows that the power-split device is successfully integrated into
vehicle model.

61
Figure 5.9: Vehicle and ICE mode comparison with powertrain components

62
Figure 5.10 shows the engine operating points, in red dots, with respect to Brake
Specific Fuel Consumption (BSFC) map. The control strategy for engine operation is
based on power request and optimum power-speed curve. Low power request and
high engine torque capacity makes the engine operation at lower side of torque
curve and not utilizing engine to its optimum capacity. Higher fuel efficiency of the
vehicle can be achieved by increasing the power-split ratio on engine side so that
engine is operated at high torque and rpm values.

Figure 5.10: Engine operating points in BSFC map of 2.9 liter engine

63
Figure 5.11 shows the comparison of MG2 operating points on efficiency map. The
MG2 operating points are in black dots. MG2 operates in orange to yellow region
giving efficiency close to 85%. Most of the torque points are low as motor capacity is
high and torque demand is comparatively low.

Figure 5.11: MG2 motor operating points on motor efficiency map

64
6. CONCLUSIONS AND FUTURE WORK

The power-split vehicle developed in Simulink environment for HIL simulation has
been implemented successfully. The power-split device allows independent engine
RPM from vehicle speed helping in engine level optimization of vehicle. The dynamic
model of third generation of THS (hybrid synergy drive) was developed and validated
using HIL simulation. An added clutch to engine carrier allows vehicle operating in
two modes. The electric vehicle mode is useful in initial acceleration and high torque
requirement. In hybrid mode, the engine performance can be improved with
variation in MG1 motor/generator and better control strategy. The HIL simulation
was validated with ControlDesk software and the results were reported.

The recommended future work includes following:

x The created model is input split configuration for hybrid synergy drive. Other
configurations of power-split can be modeled and tested.

x A better control strategy for clutched version of hybrid synergy drive can help
to improve the efficiency of the vehicle.

x Change in motor and engine rating can be done to optimize the vehicle
performance.

x An optimum gear ratio can be implemented to optimize the vehicle

performance.

x A better clutch model is necessary to incorporate the dynamics of the clutch

into the power-split model.

65
REFERENCES
1. Ehsani, M., Modern electric, hybrid electric, and fuel cell vehicles
fundamentals, theory, and design2010: Boca Raton : CRC Press, 2010.
2. Tate, E.D., M.O. Harpster, and P.J. Savagian, The Electrification of the
Automobile: From Conventional Hybrid, to Plug-in Hybrids, to Extended-
Range Electric Vehicles. SAE Int. J. Passeng. Cars - Electron. Electr. Syst.,
2008. 1(1): p. 156-166.
3. Chen, D.B., MEEM5750 Distributed Embedded Control System, 2012,
Michigan Technological University.
4. Meisel, J., An Analytic Foundation for the Toyota Prius THS-II Powertrain with
a Comparison to a Strong Parallel Hybrid-Electric Powertrain, 2006, SAE
International.
5. Aimin, D. and C. Hongwei. An analytic foundation for the Two-mode hybrid
transmission with a comparison to other hybrid vehicle power split
transmissions. in Transportation, Mechanical, and Electrical Engineering
(TMEE), 2011 International Conference on. 2011.
6. Xiaowu, Z., et al., Prius+and Volt-: Configuration Analysis of Power-Split
Hybrid Vehicles With a Single Planetary Gear. Vehicular Technology, IEEE
Transactions on, 2012. 61(8): p. 3544-3552.
7. Liu, J., Modeling, Configuration and Control Optimization of Power-split Hybrid
Vehicles, in Mechanical Engineering2007, University of Michigan.
8. Poramapojana, P., Predictive control of hybrid vehicle powertrain for
intelligent energy management, 2012, Michigan Technological University. p.
xvii, 91 p.
9. Jinming, L. and P. Huei, Modeling and Control of a Power-Split Hybrid Vehicle.
Control Systems Technology, IEEE Transactions on, 2008. 16(6): p. 1242-
1251.
10. Musardo, C., G. Rizzoni, and B. Staccia. A-ECMS: An Adaptive Algorithm for
Hybrid Electric Vehicle Energy Management. in Decision and Control, 2005
and 2005 European Control Conference. CDC-ECC '05. 44th IEEE Conference
on. 2005.
11. Namwook, K., C. Sukwon, and P. Huei, Optimal Control of Hybrid Electric
Vehicles Based on Pontryagin's Minimum Principle. Control Systems
Technology, IEEE Transactions on, 2011. 19(5): p. 1279-1287.
12. Mizuno, Y., et al., Development of New Hybrid Transmission for Compact-
Class Vehicles. SAE Int. J. Engines, 2009. 2(1): p. 742-749.
13. Willis, R., Principles of Mechanism1841, Cambridge University Press.
14. Olszewski, M., EVALUATION OF THE 2010 TOYOTA PRIUS HYBRID SYNERGY
DRIVE SYSTEM. FY2011. ORNL/TM-2010/253.
15. Kim, N., A. Rousseau, and E. Rask, Autonomie Model Validation with Test
Data for 2010 Toyota Prius, 2012, SAE International.
16. Inc., d., ASM Electric Component Reference, 2011: dSPACE GmbH,
Rathenaustraße 26, 33102 Paderborn, Germany.
17. Inc., d., dSPACE Simulator EcoLine Hardware Installation and Configuration,
2010: dSPACE GmbH, Rathenaustraße 26, 33102 Paderborn, Germany.

66