Final Report Valeo

48V HYBRIDIZATION OF A MID-SIZE VEHICLE

USING ELECTRIC MOTOR AND ELECTRIC

ASSISTED SUPERCHARGER

Powertrain Program

IFP School 2014/2015

Supervisor: S. Potteau

Authors: I.D.

Aggelos Zoufios E12792

Shixiong Zhao E12790

Stanish Gunasekaran E12593

Thomas Redlinger E12718

Date: 12th June 2015

I

ABSTRACT

In a context of growing demand for sustainable transportation worldwide, different

technical solutions for hybrid vehicles are nowadays investigated as effective ways to

improve efficiency of the driveline and thus to reduce CO2 emissions. As a matter of fact,

the CO2 emission targets set by EU (95 g/km in 2020 and 75 g/km in 2025) are extremely

demanding.

In order to reach the 2020 CO2 emission target with a spark ignition engine, solutions

needs to be developed or reinforced. The solution to be reinforced is downsizing and the

solution to be developed is hybridization. These two solutions would be able to lower CO2

emissions of spark ignition engines to at least 95 g/km.

By implementing mild hybridization (48V) in a vehicle equipped with a 1.2L downsized

turbocharged gasoline engine, it is possible to reduce fuel consumption and as a

consequence CO2 emissions comparing to the baseline diesel vehicle (Golf VII 1.6 TDI)

and meet the 2020 regulation. These gains in terms of CO2 can reach up to 10% (89 g/km).

At the same time, no compromise should be done between fuel consumption and fun-to-

drive. Turbocharged engines exhibit poor transient performance (“turbo lag”). In this

regard, some work has been done in incorporating air compressors driven by electric

machines in the air flow path of an engine. An electric supercharger allows improving fuel

consumption, turbo lag and increasing engine torque at low speed.

By implementing both an electric machine (15kW) and an electric supercharger (4kW),

the driving performance of the proposed hybrid has been maintained or even improved

reducing the turbo lag.

Finally although the price of a mild hybrid can be more expensive compared with a

conventional diesel, it has been proven that benefits can be gained back during the car

lifetime.

II

ACKNOWLEDGEMENT

We are using this opportunity to express our gratitude to everyone who supported us

through this project.

Foremost we would like to express our sincere gratitude to our supervisor, Mr Sebastien

Potteau from Valeo, for his continuous support and availability to answer any questions

we might have had. His trustful and insightful comments were appreciable all the time of

research and writing of this project.

Special thanks go out to Mr Prakash Dewagan from IFP School, for sharing his expertise

in Matlab Simulink ©.

Kind thanks also go out to Mrs Ouafae El Ganaoui-Mourlan and Mr Pascal Greau for giving

us the opportunity to work on this project.

III

TABLE OF CONTENTS

Abstract ................................................................................................................................................................ I

Acknowledgement.......................................................................................................................................... II

Introduction ....................................................................................................................................................... 1

1 Benchmarking ......................................................................................................................................... 3

1.1 Diesel .................................................................................................................................................. 3

1.2 Gasoline.............................................................................................................................................. 4

1.3 Hybrid ................................................................................................................................................. 5

1.4 Comparison and Benchmarking ............................................................................................... 7

1.5 Cycles .................................................................................................................................................. 8

1.6 Summary ........................................................................................................................................ 12

2 Engine Selection & Optimization .................................................................................................. 14

2.1 Downsizing and gear ratios .................................................................................................... 14

2.2 3-Cylinders Engine ..................................................................................................................... 17

3 Hybrid Architecture Selection........................................................................................................ 22

3.1 Degree Of Hybridization ........................................................................................................... 22

3.2 Hybrid Architecture ................................................................................................................... 23

3.2.1 Series Architecture ............................................................................................................ 23

3.2.2 Parallel Architecture ......................................................................................................... 23

3.2.3 Power Split Architecture ................................................................................................. 23

3.3 Component Sizing ....................................................................................................................... 24

3.3.1 Engine Selection .................................................................................................................. 24

3.3.2 Electrical System Selection ............................................................................................. 24

3.4 Energy Management Strategy ................................................................................................ 26

IV

3.4.1 Principle ................................................................................................................................. 26

3.4.2 Implementation .................................................................................................................. 27

3.5 Elaboration and Quantitative Rating of Possible Parallel Hybrid Topologies ..... 30

3.5.1 Conventional Architecture .............................................................................................. 30

3.5.2 Stop & Start Implementation ......................................................................................... 30

3.5.3 Mild-Hybrid Architectures .............................................................................................. 31

3.6 Final Architecture ....................................................................................................................... 35

4 Gear Selection Optimization ........................................................................................................... 37

4.1 Gear Shifting Strategy ................................................................................................................ 37

4.1.1 Principle ................................................................................................................................. 37

4.1.2 Implementation .................................................................................................................. 38

4.2 Simulations .................................................................................................................................... 38

4.2.1 NEDC Cycle ............................................................................................................................ 38

4.2.2 Real Driving Cycle In Paris .............................................................................................. 41

5 E-Supercharger Implementation .................................................................................................. 44

5.1 System Layout and Requirements ........................................................................................ 45

5.2 Simulation Methodology .......................................................................................................... 46

5.2.1 Engine Torque Calculation ............................................................................................. 46

5.2.2 Motor Torque ....................................................................................................................... 49

5.3 Performance Calculation Results .......................................................................................... 49

5.3.1 Time to Torque at Constant Engine Speeds ............................................................. 49

5.3.2 Overall Results on All Acceleration Ranges Considered ..................................... 51

5.3.3 Focus on 80-120km/h Acceleration ........................................................................... 52

6 Results ..................................................................................................................................................... 55

V

6.1 Fuel Consumption & CO2 Emissions ................................................................................... 55

6.1.1 NEDC Cycle ............................................................................................................................ 55

6.1.2 RDE Cycle............................................................................................................................... 55

6.2 System Overview ......................................................................................................................... 56

6.3 Additional Weight & Cost Estimation.................................................................................. 57

6.4 Cost to the OEM............................................................................................................................ 57

6.5 Total Cost Of Ownership Estimation ................................................................................... 58

Conclusion and Directions for further study ..................................................................................... 59

References ....................................................................................................................................................... 61

1

INTRODUCTION

Internal Combustion Engine has been the main machine for producing work and power

in the transport sector since Nicolaus Otto invented the first four stroke spark ignition

engine in 1876. Although, Otto’s engine was able in producing only a small amount of

power due to very small efficiency, huge steps have been achieved for increasing the

efficiency of an ICE (e.g. advanced materials, high pressure injection pumps, power

electronics …).

At the same time new European norms (Euro 5, Euro 6) have reinforced the limitations

for the control of the already regulated exhaust emissions coming from an ICE.

Considering an ICE, these CO2 emissions reduction can only be achieved by reducing as

much as possible the required amount of fuel to produce a certain power output, and thus

improving fuel economy.

Being the ICE inefficient enough at part loads, new techniques have been investigated in

order to eliminate the operating time of an ICE under these conditions. One solution is the

engine downsizing. That is to say forcing the engine with smaller displaced volume to

operate at higher loads and as a consequence under improved efficiency to produce the

same amount of power and torque. Another solution is the hybridization, nowadays a

promising solution for further improvement of fuel economy that contributes in the

further reduction of CO2 emissions.

The basic principle of hybridization is to implement an additional power source in the

vehicle driveline which will operate in behalf of the ICE when it has been proven to be

inefficient or in combination with the engine. The connection between the different power

sources of the vehicles can be done either in powertrain level (Parallel hybrids) or in

power level (Series hybrids). A combination between these two configurations can lead

to more complex hybrids. The main limitation as long as it concerns hybrid vehicles is the

energy source: the battery. Although battery technology has accomplished huge steps the

recent years, its incapability to store and deliver a big amount of energy is and will be

crucial for the future of the hybrid vehicles.

At the same time it is obvious that along with low fuel consumption main driver’s demand

is good performance in acceptable cost. In terms of performance, although the downsizing

2

of an engine with the use of turbocharging has been proven to be beneficial for fuel

economy, turbochargers exhibit poor transient performance especially at low engine

speeds. By implementing air compressors to the air loop, driven not by the exhaust gas

enthalpy of the turbine but by electric motors (electric superchargers), the low end torque

and as a consequence the feeling of the driver is really improved.

Regarding the economic aspect, hybrid vehicles are more expensive comparing to

conventional vehicles due to the usage of additional components (bigger battery, electric

motor, extra cables …). So the reduction of CO2 emissions with the usage of hybridization

should meet an acceptable cost threshold for both the final customer and the OEM.

The purpose of this study is to try to match or improve the fuel consumption and CO2

emissions emitted by a state of the art diesel vehicle (Golf VII 1.6L TDI) with a gasoline

hybrid vehicle without sacrificing the performance of the vehicle, and all these in an

acceptable economic package for the final customer.

3

1 BENCHMARKING

Having identified our target and our product goal, benchmarking is essential to define the

best in class performance matrices and hence define our functional requirements.

Benchmarking is the process of comparing one's business processes and performance

values to industry bests or best practices from other companies. Dimensions typically

measured are fuel consumption, performance and cost.

1.1 DIESEL

A set of various diesel cars in the C segment with comparable performance and price

ranges is compared for performance, consumption and emissions (Figure 1-1 and Figure

1-2). Considering the compromises on cost to company and the required performance

levels the top three vehicles will be shortlisted later.

The comparison provides a general overview over the range of engine power in the C

segment.

Figure 1-1 : Diesel Vehicle CO2 Emissions.

The diesel engines tend to be heavier than the gasoline counterparts while the CO2

emissions tend to be lower. However Diesel cars are generally associated with higher

soot/ smoke in the exhaust. This has led to emission norms emphasizing the installation

of DPF & after treatment devices which further increase the cost and weight of the vehicle.

1295

1395

1185

1385

99 99

84

104

80

90

100

110

120

130

140

150

800

900

1000

1100

1200

1300

1400

1500

Golf VII 1,6 TDI BMW 116dEfficient

DynamicsEdition

Peugeot 308Blue HDI 120

Alfa RomeoGiullieta 1,6JTDM-2 105

CO

2 e

mis

sio

n (

g/km

)

Cu

rb w

eig

ht

(kg)

Curb weight CO2 emission

4

Figure 1-2 : Diesel Vehicle Performances.

1.2 GASOLINE

A comparison of similar gasoline cars with respect to the Golf VII are selected for

benchmarking. Some of the performance indicators considered are: stand-still

acceleration (0-100km/h), roll-on acceleration (80-120km/h) and maximum speed

(Figure 1-4).

A common trend observed is that the average fuel consumption and CO2 emission levels

are higher than for the diesel engines (Figure 1-3). And observing the market trends

towards the gasoline vehicles it is a necessity to implement suitable strategies to reduce

these levels. These will be discussed later in this report.

The weight range is between 1050 kg (Ford Fiesta) to 1350 kg in the case of Peugeot 207.

All the engines are however front transversal and weight to power ratio lies between 55

– 60 kg/kW for the analyzed set of cars. Gasoline cars own a huge portion of the market

share outside European market and the European legislations show a proclivity to

gasoline cars with the progressing emission trends. Hence choosing a gasoline version of

the hybrid makes more sense with the legislation and technology trend.

189

195193

185

11

9,4

9,7

11,3

9

9,5

10

10,5

11

11,5

12

175

180

185

190

195

200

Golf VII 1,6 TDI BMW 116dEfficient

DynamicsEdition

Peugeot 308Blue HDI 120

Alfa RomeoGiullieta 1,6JTDM-2 105

0-1

00

(se

c )

Max

Sp

ee

d (

Km

/hr)

5

Figure 1-3 : Gasoline Vehicles CO2 Emissions.

Figure 1-4 : Gasoline Vehicles Performances.

1.3 HYBRID

The hybrid vehicles have smaller engines which are generally downsized and this

contributes to a lower emission level (Figure 1-5). The series and parallel hybrid

architectures are very popular among manufacturers and between both the parallel

architecture require smaller battery pack and is more suitable for mild hybrid application

1090

1200

1160

1205

110114

129

115

80

90

100

110

120

130

140

150

1000

1050

1100

1150

1200

1250

1300

Peugeot 308 II1.2 THP 130 PS

(2014)

Ford Focus III1.0 EcoBoost125 PS (2012)

Nissan Juke 1.2DIG-T 115 PS

(2014)

Hyundai i20 1.4100 PS (2015)

CO

2 e

mis

sio

ns

(g /

km)

Cu

rb w

eig

ht

(kg)

Curb weight CO2 emission

207

192

178

184

11

13,2

15,1

8,5

2

4

6

8

10

12

14

16

175

180

185

190

195

200

205

210

Peugeot 308 II1.2 THP 130 PS

(2014)

Ford Focus III 1.0EcoBoost 125 PS

(2012)

Nissan Juke 1.2DIG-T 115 PS

(2014)

Hyundai i20 1.4100 PS (2015)

80

-12

0 (

sec)

Max

Sp

ee

d

(Km

/hr)

800-120( sec) 80-120 acc

6

for a C segment vehicle. Parallel architecture allows thermal and electrical sources to

provide torque simultaneously and improve engine efficiency.

Figure 1-5 : Hybrid Vehicles CO2 Emissions.

Figure 1-6 : Hybrid Vehicles Performances.

The drawbacks however of the hybrid vehicle are as follows:

High Prices and Vehicle weight (battery pack)

Highway Consumption not improved

Difficulty when driving aggressively, lack of dynamism in acceleration, engine

reacts late on highway.

12461310

1370

1660

1295

101

84 8288

102

80

90

100

110

120

130

140

150

160

170

800

900

1000

1100

1200

1300

1400

1500

1600

1700

HondaInsight II

Toyota Auris136h

Lexus CT200h

Peugeot3008 Hybrid

Golf VII 1,6TDI

CO

2 e

mis

sio

n (

g/k

m)

Cu

rb w

eig

ht

(kg)

Curb weight CO2 emissions

182180 180

185

192

11,1

8,2 8,1

6,6

9,6

0

2

4

6

8

10

12

170

175

180

185

190

195

200

205

210

HondaInsight II

Toyota Auris136h

Lexus CT200h

Peugeot3008 Hybrid

Golf VII 1,6TDI

80

-12

0 a

cc (

sec)

Max

Sp

ee

d

(Km

/hr)

Max Speed 80 - 120 Km/hr

7

1.4 COMPARISON AND BENCHMARKING

Here are few noteworthy points to note:

• Gasoline cars have higher CO2 emissions hence our choice to hybridize a gasoline

vehicle as a market replacement of the Golf VII is interesting.

• The important customer expectations from the C segment car are observed to be

pick up acceleration, roll on acceleration, max speed and vehicle size.

• The conclusion from the benchmarking study is shown in Figure 1-7.

Figure 1-7 : Fuel Consumption (top) and CO2 Emissions (bottom) Summary.

3,9 3,8

4,8

3,23,4

5

44,4

5,4

0

1

2

3

4

5

6

Diesel Hybrid Gasoline

Fue

l co

nsu

mp

tio

n (

L/1

00

km)

Overall Fuel Consumption Comparison

99

84

110

84 88

114104 101

124

0

20

40

60

80

100

120

140

Diesel Hybrid Gasoline

CO2 Comparison

8

1.5 CYCLES

An emission test cycle is a protocol contained in an emission standard to allow consistent

comparable measurement of emissions for different engines. Test cycles specify the

conditions under which the engine is operating during the emission test.

Light duty vehicles and motorcycles are tested on a chassis dynamometer test bench using

standardised vehicle speed cycles to test for compliance with the exhaust emissions and

CO2 emissions. There are various legislations adopted throughout the world depending

on the driving conditions, political and technological scenario.

The various cycles today include:

FTP – Commonly known as the FTP 75 defined by the US EPA is a set of regulations to

measure the tailpipe emission and consumption of cars. The characteristics of the

cycle are:

Distance travelled: 17.77 km (11.04 miles)

Duration: 1874 seconds

Average speed: 34.1 km/h (21.2 mph)

Figure 1-8 : FTP Cycle.

JC08 – Japanese driving conditions are represented through JC08 chassis

dynamometer test cycle for light vehicles (< 3500 kg). The test represents driving in

congested city traffic, including idling periods and frequently alternating acceleration

and deceleration. Measurement is made twice, with a cold start and with a warm start.

The test is used for emission measurement and fuel economy determination, for

gasoline and diesel vehicles.

9

Figure 1-9 : JC08 Cycle.

NEDC – New European Driving cycle is supposed to represent the real driving

conditions in Europe (1997). It consists of four cycles of ECE -15 (urban) mode and

one EUDC (Extra urban or highway mode). The cycle is tested at 20-30°C conditions

on a roller test bench to represent flat road, and zero wind conditions. The cycle is

however criticized for not representing real motorway conditions and not necessarily

repeatable and comparable leading to cycle beating.

Figure 1-10 : NEDC Cycle.

WLTP - The Worldwide harmonized Light vehicles Test Procedures (WLTP) defines a

global harmonized standard for determining the level of pollutants and CO2

emissions, fuel consumption, and electric range from light-duty vehicles around the

globe including European Union, Japan, India and China under the UNECE vehicle

regulations. The cycle imposes strict constraints regarding road load (motion

resistance), gear shifting, total car weight (by including optional equipment, cargo and

10

passengers), fuel quality, ambient temperature, and tire selection and pressure. The

cycle is applied based on vehicle class which is derived from the car power to weight

ratio.

Figure 1-11 : WLTC Cycle.

ARTEMIS - The Common Artemis Driving Cycles are chassis dynamometer procedures

developed within the European Artemis project, based on statistical analysis of a large

database of European real world driving patterns. The cycles include three driving

schedules: Urban, Rural road and Motorway. The Motorway cycle has two variants

with maximum speeds of 130 and 150 km/h. Artemis cycle definitions also include

gear shifting strategies.

Figure 1-12 : Artemis Cycle.

The pros and cons of all the major cycles are enlisted in Table 1-1. It can be observed that

each cycle has its unique characteristic, very specific to the driving needs of the area it

represents. The choice of cycle is very important to fix realistic target model, in a given

driving condition. NEDC cycle though a common European benchmark is very steady state

11

and does not represent real driving conditions as shown later in this report. However it

must be taken in the stride of setting a common emission limit and also since it provides

a scale of comparison with the Volkswagen Golf VII Diesel engine which has emission data

based on the NEDC.

Pros Cons Remarks

NEDC Common European

benchmark

Not realistic Constant acceleration,

deceleration, speed

Cycle beating

No auxiliary power

consumption

WLTP Gear shifting, Weight

accounted May vary from the local conditions

DATA from 5 different countries

and 3 different segment cars Optimal shift points

Artemis Includes gear shifting Not used for certification

of pollutants or consumption

Statistical model

JC08 Very transient Does not represent

normal driving Congested city

traffic

Aggressive driving and air-conditioning included with US06

high speed and motorway

Still not completely representative of European driving

FTA

Table 1-1 : Comparison of Cycle Benefits.

This report shows that the Real driving emissions are far different from the Cycle chosen.

The upcoming EU6c Emission Regulation will implement RDE as an additional

requirement in the 2017 - 2020 timeframe. Compared to the current test environments,

which are designed and optimized for perfect reproducibility and a removal of external

influences, driving a vehicle on the road under "real-life" conditions will never be 100%

reproducible. The influence of the road profile, the ambient conditions, the traffic

situation and the behaviour of the driver itself will significantly influence the results. One

to one comparison of test results will not be possible, but it is crucial from the point of

view of real world global limits for pollutants. A summary of improvement in performance

due to hybridisation in the RDE cycle is also included in the report.

12

1.6 Summary

Figure 1-13 : Trend showing reduced fuel consumption.

The customers represent a huge diversity in the type of driving profile and the conditions

they drive in. Sustained efforts need to be taken to reduce the overall fuel consumption

whilst catering to customer demands (keeping fun to drive features, etc). This balance is

determined by developing a cycle that is robust and realistic. The different demands and

driving conditions translate to different energy management strategies and powertrain

utilisation as shown below. The overall trend has been to reduce the fuel consumption

and work is underway for OEMS and Suppliers.

182 g/km

172 g/km

Figure 1-14 : Different Utilization of Engine Energy Management in Different Cycles.

13

Figure 1-15 : Cycles Summary.

14

2 ENGINE SELECTION & OPTIMIZATION

2.1 DOWNSIZING AND GEAR RATIOS

The engine selected for this study is a 1.6L turbocharged GDI engine. In order to reduce

the fuel consumption, the engine was further downsized to 1.2L. In order to downsize the

engine from 1.6L to 1.2L the specific torque was used:

𝑇 = 𝑏𝑚𝑒𝑝 ∗ 𝑉𝑑

4 ∗ 𝜋 (Equation 2-1)

By checking (Equation 2-1 it is easy to understand that in order to obtain the same torque

from the 1.2L engine comparing to 1.6L, it is necessary for the engine to operate at higher

bmep (higher load) and as a consequence in areas with higher efficiency and lower brake

specific fuel consumption (bsfc). That is the main principle of downsizing.

The full load curve of the engine in terms of specific torque and specific power can be seen

in Figure 2-1.

The maximum torque and power delivered by the engine are respectively 205 Nm@1750

rpm and 80.7 kW@5500 rpm.

Figure 2-1 : Full Load Curve of the Engine Using Specific Torque and Specific Power.

In order to have the desired performance of the vehicle in terms of maximum vehicle

comparison with the reference vehicle of our study (Golf VII 1.6L TDI), the gear ratios

0,0

10,0

20,0

30,0

40,0

50,0

60,0

70,0

80,0

0

20

40

60

80

100

120

140

160

180

0 1000 2000 3000 4000 5000 6000 7000

Po

wer

[K

W]

Torq

ue

[Nm

]

Engine Speed [RPM]

Specific Torque (Nm/l) Specific Power (kW/l)

15

have to be selected. There are two main resistive forces applied to a vehicle during its

movement.

Rolling Resistance Force

𝐹𝑡𝑦𝑟𝑒 [𝑁] =𝑀 [𝑘𝑔] ∗ 𝑔 ∗ 𝐶𝑟𝑟 [

𝑘𝑔𝑡 ]

1000 (Equation 2-2)

Aerodynamic Force

𝐹𝑎𝑒𝑟𝑜 [𝑁] =1

2∗ 𝜌 [

𝐾𝑔

𝑚3] ∗ 𝑆 [𝑚2] ∗ 𝐶𝑥 ∗ 𝑉2 [

𝑚2

𝑠2] (Equation 2-3)

In our study the third resisting force due to the slope of the road is considered equal to

zero. So the sum of these two forces will give us the total resisting force (𝐹𝑟𝑒𝑠𝑖𝑠𝑡 [𝑁] =

𝐹𝑡𝑦𝑟𝑒 [𝑁] + 𝐹𝑎𝑒𝑟𝑜 [𝑁]). The resistive power the vehicle has to overcome is given by the

following equation:

𝑃𝑟𝑒𝑠𝑖𝑠𝑡 [𝑊] = 𝐹𝑟𝑒𝑠𝑖𝑠𝑡 [𝑁] ∗ 𝑉 [𝑚

𝑠] (Equation 2-4)

When the power that is produced by the engine multiplied by the transmission efficiency

is equal to the maximum resisting power applied to the vehicle the vehicle has obtained

its maximum speed. Then the last gear ratio can be deduced by looking at the ratio

between the highest possible vehicle speed and the maximum power engine speed. The

final gear ratio is computed using the following formula:

𝐿𝑖 [

𝐾𝑚ℎ

1000𝑟𝑝𝑚] = 𝑉1000 =

𝑉𝑒ℎ𝑖𝑐𝑙𝑒 𝑆𝑝𝑒𝑒𝑑 [𝐾𝑚/ℎ]

𝑁 [𝑟𝑝𝑚]/1000 (Equation 2-5)

This last gear ratio is optimal regarding maximum vehicle speed.

Three types of gearbox were tested. The optimal one, a short one (-10% of the optimal

one) and a long one (+10% of the optimal one). The maximum vehicle speed with these

three types of gearbox is shown in Figure 2-2. The maximum vehicle speed with the

optimal, short and long gearbox are respectively 190, 171 and 188 km/h. These three

types of gearbox were also tested in order to obtain the roll-on acceleration from 80-120

km/h. The results can be seen in Figure 2-3. The time required with the longest gearbox

16

to reach 120 km/h is 12 s. This performance is close to the targeted one (11.6s) and since

a long gearbox is better considering fuel consumption, this gearbox with a final gear ratio

of 38 has been selected.

Figure 2-2 : Maximum Vehicle Speed for the three Gearboxes (Optimal, Short and Long).

Figure 2-3 : Roll-on Acceleration from 80-120 km/h for the three Gearboxes (Optimal, Short and Long).

0

10

20

30

40

50

60

70

80

0 50 100 150 200 250 300

Po

wer

(kW

)

Vehicle Speed (km/h)

Maximum Vehicle SpeedPresist (kW) optimal gearbox Short Gearbox Long Gearbox

80

85

90

95

100

105

110

115

120

125

0 2 4 6 8 10 12 14

Veh

ivle

Sp

eed

(km

/h)

Time (s)

Vehicle Speed = f(time)

17

After the selection of the final gear ratio, it is important to obtain the gear ratios associated

to the other gears. The deduced gear ratios are shown in Table 2-1.

Gear Number Gear ratio

1st 5.24

2nd 2.89

3rd 1.89

4th 1.33

5th 0.97

Table 2-1 : Gear Ratios.

It has to be noticed that the first gear ratio has been set to reach a take of performance of

4.4 m/s2. Then the intermediate ratios have been selected using a combined steeping

(average between geometric and arithmetic stepping).

2.2 3-CYLINDERS ENGINE

Heat transfer affects engine performance, efficiency and emissions. For a given mass of

fuel within the cylinder, an increase in heat transfer to the combustion chamber walls will

lower the average combustion gas temperature and pressure. This will lead in a work per

cycle transferred to the piston reduction. Thus specific power and efficiency are affected

by the magnitude of the engine heat transfer.

Heat transfer between the unburned charge and the chamber walls in spark–ignition

engines affects the onset of knock which, by limiting the compression ratio, also influences

power and efficiency. Moreover changes in gas temperature due to the heat transfer

impact the emission formation processes, both within the engine’s cylinder and in the

exhaust system where afterburning of CO and HC occurs. The exhaust temperature also

governs the power that can be obtained from exhaust energy recovery devices such as a

turbocharger turbine.

Friction is also affected by engine heat transfer and contributes to the coolant load. The

cylinder liner temperature governs the piston and ring lubricating oil film temperature,

18

and hence its viscosity. Some of the mechanical energy dissipated due to friction must be

rejected to the atmosphere by the cooling system. The fan and water pump power

requirements are determined by the magnitude of the heat rejected. Thus the importance

of engine heat transfer is clear. The energy balance in an engine can be seen in Figure 2-4.

Figure 2-4 : Energy Balance in an Internal Combustion Engine.

Generally the fuel energy released during combustion in an ICE is converted to the

following forms with the following distribution:

Mechanical work in the crankshaft ~ 30%

Heat losses in the exhaust ~ 30%

Coolant losses in the coolant ~ 30%

Miscellaneous losses (radiation, free convection, oil if separately cooled) ~10%

Nf = Ne + Ncool + Nexh + Nmisc (Equation 2-6)

The heat flux into the wall has in general both a convective and a radiative component.

The heat flux is conducted through the wall and then convected from the wall to the

coolant. Heat is transferred by forced convection between the in-cylinder gases and the

cylinder head, valves, cylinder walls, and piston during induction, compression, expansion

and exhaust processes. Heat is transferred by forced convection from the cylinder walls

and head to the coolant, and from the piston to the lubricant or other piston coolant.

Substantial convective heat transfer occurs to the exhaust valve, exhaust port, and exhaust

Air

Fuel

Exhaust losses (Nexh)

Coolant losses (Ncool) Different losses

(Nmisc)

ICE

Mechanical Work (Ne)

19

manifold during the exhaust process. However the biggest amount of heat transfer takes

place between the burned gases and the cylinder wall during the phase of combustion and

expansion [1]. The amount of heat losses to the liner is given by the (Equation 2-7.

�̇� = ℎ ∗ 𝐴 ∗ (𝑇𝑔 − 𝑇𝑤𝑎𝑙𝑙) (Equation 2-7)

In the (Equation 2-7, h is the heat transfer coefficient, A the surface of the liner being in

contact with the coolant, Tg the gas temperature and Twall the temperature of the liner.

It can be seen that the amount of heat rate is strongly dependent on the available area of

the cylinder liner. By reducing this area a benefit considering the heat losses could be

obtained.

By reducing the number of cylinders from 4 to 3 it is possible to reduce the available area

for heat transfer and as a consequence the amount of heat losses in the coolant. The basic

assumptions made for going from 4 to 3 cylinders are:

Exhaust losses remain constant at 30% of the total fuel energy

Stroke to Bore ratio (S/B) for conventional gasoline engines is 0.8

Constant mean wall temperature

Bmep unchanged

According to these assumptions any possible gain in terms of reduction of the available

heat transfer area and as a consequence a reduction of the heat losses will be in the benefit

of the brake specific fuel consumption of the engine. Indeed for the engine to maintain the

same load (bmep), with reduced heat losses the amount of air and fuel introduced in the

cylinder is lower.

The displaced volume of the engine is 1.2L. Displaced volume can be calculated from the

(Equation 2-8.

𝑉𝑑 =𝑝𝑖 ∗ 𝐵2

4∗ 𝑆 (Equation 2-8)

Where B is the piston diameter (Bore) and S the stroke of the engine.

20

Using the previous assumption and solving the (Equation 2-8 for the bore it is possible to

obtain a value for the piston diameter and then the stroke. The available area for heat

losses is given by the following equation:

𝐴 = 𝜋 ∗ 𝐵 ∗ 𝑆 (Equation 2-9)

The results and the benefit of the reduction in heat transfer area are below shown in Table

2-2.

4 cylinders 3 cylinders Percentage of area

reduction

S

[cm]

B

[cm]

Vd

[cm3]

A

[cm2]

S

[cm]

B

[cm]

Vd

[cm3]

A

[cm2] [%]

6.25 7.81 1200 614.3 6.88 8.6 1200 558 9.16

Table 2-2 : Difference in Heat Transfer Area Obtained Between 4 and 3 Cylinders.

According to [1], although the heat losses are such a substantial part of the fuel energy

input, elimination of heat losses would only allow a fraction of the heat transferred to the

combustion chamber walls to be converted to useful work. The remainder would leave

the engine as sensible exhaust enthalpy. Considering an automotive high – speed naturally

aspirated CI engine with a compression ratio of 15. The indicated efficiency is 45%, and

25 percent of the fuel energy is carried away by the cooling water. Of this 25%, about 2%

is due to the friction. Of the remaining 23%, about 8% is heat loss during combustion, 6%

heat loss during expansion, and 9% heat loss during exhaust. From the 8% lost during

combustion about half (4% of the fuel energy) could be converted into useful work on the

piston. From the 6% heat loss during expansion, about one – third (2%) could have been

utilized. Thus of the 25% lost to the cooling system, only about 6% could have been

converted to useful work on the piston, which would increase the indicated efficiency of

the engine from 45 to 51%. (Figure 2-5)

21

Figure 2-5 : Energy Flow Diagram for an Internal Combustion Engine. [1]

For a spark ignition engine, the conversion to useful work will be lower, because the

compression ratio is lower. However, the heat losses at part load (which constitute an

important operating regime for automobile use) are a substantially larger fraction of the

fuel heating value. Studies with computer simulations of the SI engine operating cycle

indicate that at typical part–load conditions a proportional reduction in combustion

chamber wall heat losses of 10% results in a proportional increase (improvement) in

brake specific fuel conversion of about 3 percent [1].

By subtracting the previously calculated 9.16% reduction in the available heat transfer

area a 2.75% improvement in bsfc is obtained.

22

3 HYBRID ARCHITECTURE SELECTION

3.1 DEGREE OF HYBRIDIZATION

Hybridization represents undoubtedly a valuable option to improve fuel savings thanks

to the following efficiency enhancements [2]:

Deleting idling losses by switching off the engine when vehicle stops,

implementing a Stop & Start strategy.

Using regenerative braking in order to recover part of the vehicle kinetic energy

during deceleration instead of dissipating it through braking system.

Operating the internal combustion engine closer to its best efficiency, trying

to avoid its use under highly inefficient operating conditions, thanks to the

additional degree of freedom provided by the electrical power source and energy

storage devices.

Enabling internal combustion engine downsizing while still maintaining

acceptable vehicle performance thanks to the additional boosting which can be

provided by the electric power source.

There are different degree of hybridization: Micro, Mild, Full and Plug-in hybrids.

The Plug-in hybrids are out of the scope of this study. The Micro-Hybrid cars are only able

to save fuel consumption through stop and start implementation (~5% fuel savings). The

Full-Hybrid vehicles such as the Toyota Prius utilize expensive high capacity motor-

generator integrated alongside the drive train and high capacity battery packs (more than

48V – high voltage lead to safety measures) [3]. Mild-Hybrid systems have the benefit of

incorporating fewer changes in the system architecture in comparison to Full-Hybrid

architecture and consequently being more ready for conversion from a conventional to a

hybrid powertrain. In addition it offers the following advantages over strong hybrids [3]:

A lower cost to benefit ratio

Higher power to weight ratio

Easy mechanical integration on production vehicles.

Lower risk factor owing to lower operating voltage.

No requirement for a complex energy management control strategy.

23

Thus it appears clearly that the Mild-Hybrid topology is the most suitable for this study.

3.2 HYBRID ARCHITECTURE

It exists, three different ways to arrange a hybrid architecture: series, parallel and power

split architecture.

3.2.1 SERIES ARCHITECTURE

In this kind of architectures, all of the energy from the engine goes through the electric

system before reaching the wheels. This allows the engine to operate on maximum

efficiency point but requires high energy storage capacity (large batteries and therefore

high cost). Also this architecture induces additional electric losses through the whole

electric system. [5]

3.2.2 PARALLEL ARCHITECTURE

In this architecture the propulsion at the wheel can be ensured by either the engine or the

electric system or both. The required size of the electric system is thus smaller reducing

the cost.

Moreover unlike the series hybrid, the energy from the engine does not have to go through

the electric system lowering losses when the engine is operating efficiently. But control of

engine operating points is more difficult than in a series architecture. This architecture

has been applied in the Honda hybrids. [5]

3.2.3 POWER SPLIT ARCHITECTURE

This architecture combines features of the two previous ones. One part of the engine’s

power is connected to the wheels through a series branch (power going through electric

system), and a second part through a parallel branch (engine power going directly to the

wheels). The series branch allows controlling the engine more efficiently. The size of the

batteries and electric losses are, however, kept relatively small compared with a series

architecture thanks to the parallel branch. The most known application of the power split

architecture are the Toyota hybrids. [5]

In comparison to parallel hybrids the battery size for a power split hybrid is larger at iso-

total peak power due to electrical losses, leading to an increase in price.

24

As the parallel hybrid require fewer changes when converting from a conventional to a

hybrid powertrain and lower cost thanks to smaller battery, the parallel architecture has

been chosen for this study.

3.3 COMPONENT SIZING

3.3.1 ENGINE SELECTION

The vehicle propulsion being assisted by an electric motor it is then possible to size the

engine smaller than normally required. In this way the engine will be loaded closer to its

maximum efficiency points (close to maximum torque curve). The downsizing will lead to

a fuel consumption improvement thanks to better efficiency without compromising the

fun to drive since the E-Machine will assist the propulsion during acceleration.

The E-Machine used in a Mild-Hybrid system is typically more powerful than a

conventional 12V starter and can quickly start the engine without the driver experiencing

an uncomfortable delay. The system efficiency can be further improved by charging the

batteries when the engine is operating at low load condition. [3]

3.3.2 ELECTRICAL SYSTEM SELECTION

The electric motor power capacity ranges from 1.5kW to 15kW while operating voltage

of a mild hybrid systems vary from 12volts to 64volts. [3]

The characteristics of the E-Machine used to quantitatively rate the possible hybrid

architecture are summarized in Table 3-1.

Rated Power (kW) 15

Max Torque (Nm) 70

Max Speed (RPM) 18 000

Efficiency 0.9

Table 3-1 : E-Machine Characteristics.

To size the electric motor a sensitivity study is conducted by using the machine map

(Figure 3-1) and adapting the maximum torque. Although the efficiency is engine speed

dependent it is assumed constant and equal to 90%.

25

Figure 3-1 : E-Machine Efficiency Map. [2]

Regarding the battery, the voltage is one of the constraint of the project: it is fixed to 48V.

The batteries are sized for power. More specifically, the output power of the batteries is

the peak power of the motor divided by the average motor efficiency. Thus In this case the

battery pack power should be at least 16.7kW. A combination of 17 Ultra-High Power Cell

(characteristics summarized in Table 3-2) in series is used to quantitatively rate the

possible hybrid architecture providing 18.9kW.

Nominal Capacity (Ah) 15

Mass (kg) 0.440

Nominal Voltage (V) 2.7

Maximum Power (W) 1110

Energy (Wh) 55.5

Charge-Discharge Efficiency (%) 100

Table 3-2 : Ultra-High Power Cell Characteristics. [4]

A pack of high power cell battery has been chosen since in this study the aim is to have

high power available during short periods (e.g. use of electric supercharger, engine

cranking by E-Machine).

The previously optimized engine and the previously described E-Machine give a

hybridization ratio of:

𝑅ℎ 𝑝𝑎𝑟𝑎𝑙𝑙𝑒𝑙 =𝑃𝐼𝐶𝐸

𝑃𝐼𝐶𝐸 + 𝑃𝑒𝑙𝑡= 84% (Equation 3-1)

26

3.4 ENERGY MANAGEMENT STRATEGY

3.4.1 PRINCIPLE

The degree of freedom of a parallel HEV is the engine torque (Te), this architecture

combining mechanical power through a mechanical node (Figure 3-2). The energy

management strategy has thus to select the engine torque that will minimize the fuel

consumption, this implies using the battery as a buffer. As shown on Figure 3-2, the

gearbox ratio can also be set as a degree of freedom, this will be investigated later in the

gear optimization section.

Figure 3-2 : Parallel HEVs Degree of Freedom. [6]

Let’s introduce now the Hamiltonian function (H) and the equivalence factor (S0):

𝐻 = 𝑃𝑓 + 𝑆0 × 𝑃𝑒𝑐ℎ (Equation 3-2)

Where:

𝑃𝑓 = 𝐻𝐿𝐻𝑉 × 𝑚𝑓̇ is the fuel power,

𝑃𝑒𝑐ℎ = −𝑆𝑂𝐶̇ × 𝑈𝑜𝑐 × 𝑄𝑏 is the electrochemical power,

𝑈𝑜𝑐 is the open-circuit voltage of the battery,

𝑄𝑏 is the battery capacity,

S0 is a non-dimensional parameter. [6]

This function allows optimizing the overall fuel consumption by introducing an equivalent

consumption for the power taken from the battery. It estimates how much fuel

correspond to the battery consumption. As at any time the Hamiltonian function depend

27

on the degree of freedom (engine torque) the optimal engine torque regarding fuel

consumption is the one that minimizes H.

This strategy generally allows engine operating points either to be suppressed

(powertrain goes full electric for cranking or to avoid inefficient low load points) or

moved to better efficiency points (recharge mode when battery is depleting, boost mode)

(Figure 3-3).

Figure 3-3 : Energy Management Strategy. [5]

NB: The function to crank the engine via the E-Machine will only be implemented on the

optimized energy management strategy (see section 4).

3.4.2 IMPLEMENTATION

The energy management strategy is implemented via Matlab/Simulink ©. As previously

described the aim of this submodel is to compute the engine and motor torque to generate

the required torque at the wheels while minimizing fuel consumption. It consists in

defining an array of possible engine torque candidates and select the one that minimizes

the Hamilton function seen in the previous section while respecting the following

constraints:

Engine speed ⋲ [idle speed, max speed]

Engine torque ⋲ [friction torque, max torque = f(RPM)]

Motor speed ⋲ [min speed, max speed]

Motor torque ⋲ [max regenerative torque, max motoring torque]

The architecture chosen cannot be recharged from the grid (not plug-in), it will

therefore operate in a charge sustaining mode: 𝑆𝑂𝐶𝑓𝑖𝑛𝑎𝑙 = 𝑆𝑂𝐶𝑖𝑛𝑖𝑡𝑖𝑎𝑙

Battery SOC ⋲ [min SOC, max SOC]

28

The input of this submodel are the vehicle speed, gear index, powertrain required torque

and the battery state of charge. (Figure 3-4)

Figure 3-4 : Energy Management Submodel.

The gear index and the vehicle speed are fixed by the cycle and the torque demand is

estimated as follow:

𝑇𝑃𝑊𝑇 = (𝐹𝑎𝑒𝑟𝑜 + 𝐹𝑡𝑦𝑟𝑒 + 𝐹𝑠𝑙𝑜𝑝𝑒) × 𝑅𝑡𝑦𝑟𝑒 +𝐽 × 𝑎𝑣𝑒ℎ

𝑅𝑡𝑦𝑟𝑒 (Equation 3-3)

Where :

𝐹𝑎𝑒𝑟𝑜 =1

2𝜌𝑆𝐶𝑥 × 𝑉2 is the aerodynamic force,

𝐹𝑡𝑦𝑟𝑒 = 𝑀 ∗ 𝑔 ∗𝐶𝑟𝑟

1000 is the rolling resistance force,

𝐹𝑠𝑙𝑜𝑝𝑒 = 𝑀 ∗ 𝑔 ∗ sin (𝛼) is the slope resistance force.

SCx (m²) 0.66

(kg/m3) 1.225

M (kg) 1295

Crr (kg/t) 8.5

Table 3-3 : Vehicle Characteristics.

The main outputs of the submodel are the engine and motor torque that minimize the fuel

consumption. The sum of these torque should equals the powertrain required torque at

the mechanical node that link the engine and the E-Machine. (Figure 3-2)

The array of possible candidates for the engine torque is composed of 10 values equally

spaced between the min and the max possible engine torques at the current engine speed

plus one value corresponding to a purely ICE mode corresponding to the actual gear ratio

29

and a 0 to represent the full electric mode. Thus there are 12 possible engine torques.

Knowing the required torque at the wheel (input) it is then possible to deduce the 12

corresponding value of the motor torque taking into account the gear ratio of the electric

pathway.

Figure 3-5 : Engine (left) and Motor (right) Torque Candidates on NEDC Cycle as a Function of Time.

From the previous torques calculations it is then possible to estimate the Hamiltonian

function by evaluating the fuel power and electrochemical power. To prevent any issue

with the constraints previously set unfeasibility flags have been implemented to identify

which of the solution candidates are not admissible. Finally all that remains is to find the

candidate that minimizes the Hamiltonian and that has not raised any unfeasible flag.

The model used for the E-Machine, thermal engine and battery were provided during a

supervised classwork on hybrid management strategy. [6]

In order to satisfy the charge sustaining mode the equivalent factor S0 has to be tuned.

The optimal equivalence factor is chosen considering the following algorithm:

Figure 3-6 : Flowchart of the offline Energy Management Strategy. [6]

0 200 400 600 800 1000

0

50

100

150

200

Time [s]

Ten

g [N

m]

Engine Torque candidates

0 200 400 600 800 1000-80

-60

-40

-20

0

20

40

60

80

Time [s]

Tm

ot [

Nm

]

Motor Torque candidates

30

3.5 ELABORATION AND QUANTITATIVE RATING OF POSSIBLE PARALLEL

HYBRID TOPOLOGIES

In order to evaluate the potential benefit of such architectures it is necessary to assess the

fuel consumption of the conventional gasoline powertrain. Comparison between different

architectures are based on NEDC cycle.

3.5.1 CONVENTIONAL ARCHITECTURE

For purely ICE powertrain there is no energy management strategy. Indeed the engine

torque has to provide the torque required at the wheels; there is no degree of freedom.

The thermal engine chosen for this study is a 3-cylinder 1.2L gasoline engine with direct

injection. It is assumed that 10% of crankshaft output power is lost due to friction and

auxiliaries. The transmission is a 5-gear manual transmission without any mechanical

loss.

Figure 3-7 : Engine Operating Points on NEDC (Conventional Engine).

The fuel consumption on NEDC cycle for this particular engine is 5.38 L/100km.

Considering a gCO2/km ⇄ L/100km conversion coefficient of 23.81 [7] this engine emits

128 gCO2/km, this is far above the target emission of the diesel Golf VII (99 gCO2/km).

3.5.2 STOP & START IMPLEMENTATION

This is an ideal Stop & Start which only considers that fuel consumption is 0 when vehicle

is at idle. Thus this architecture has the same engine operating point as the conventional

1000 2000 3000 4000 5000 6000

0

50

100

150

200

w [RPM]

T [

Nm

]

Engine Operating Points

31

one (Figure 3-7). The consumption on NEDC cycle is 5.09 L/100km, thus Stop & Start

implementation can achieve a fuel benefit of 5.4% and emits only 121 gCO2/km. But as

expected this reduction of CO2 emissions is not enough. It is mandatory to implement a

Mild-Hybrid architecture.

3.5.3 MILD-HYBRID ARCHITECTURES

In this section is investigated the potential benefits of the different parallel hybrid

architectures (P1, P2, P3 and P4, described in Figure 3-8)

Figure 3-8 : Different Parallel Hybrid Architecture. [8]

The energy strategy management previously described and the battery pack and electric

machine previously sized will be used to quantitatively rate the fuel savings and emission

reduction. To simulate the 4 different architectures the global architecture displayed in

Figure 3-9 is employed.

Figure 3-9 : Global Hybrid Architecture. (Similar than HOT Software)

The gear ratios Rg, Rn, Rc, Rm and Rd are used to simulate clutches and transmission

ratios. Rg is always set to 0 since it is only useful to simulate a series hybrid architecture.

Rn allows simulating a clutch between the engine and the electric machine which is typical

of a P2 architecture. Rc represents the ratio of the manual transmission and Rm the ratio

between the e-machine shaft and the manual transmission output shaft (typical of P3

architectures with double shaft). Finally Rd is the final drive ratio.

32

For all the architecture tested it is assumed that:

200W electric power is dedicated to ECU and auxiliaries.

All the braking torque can be recovered by the electric machine.

Engine resistive torque (when run by e-Machine) is 5Nm.

No use of electric machine power for cranking.

The gear shifting is imposed on NEDC cycle.

3.5.3.1 P1 ARCHITECTURE

The first architecture to be tested is the P1 architecture. (Figure 3-10)

Figure 3-10 : P1 Hybrid Architecture.

As shown on Figure 3-10 the Motor (M) is directly linked to the ICE, so it is not possible

to recover energy when the driver is pressing the clutch pedal. In this way the E-Machine

can only run full electric and regenerate energy during braking when the clutch is

engaged, so the E-Machine will undergo the engine resistive power due to friction. But as

no additional transmission is required the extra-weight implied by the hybridization is

limited. Although this configuration has obvious drawbacks the fuel consumption is

improved since the bad efficiency points at low load are removed (see green box on Figure

3-11) compared to the conventional gasoline engine (Figure 3-7).

Figure 3-11 : Engine and Electric Machine Operating Points on NEDC (P1 Architecture).

1000 2000 3000 4000 5000 6000

0

50

100

150

200

w [RPM]

T [

Nm

]

Engine Operating Points

0 2000 4000 6000 8000 10000-80

-60

-40

-20

0

20

40

60

80

w [RPM]

T [

Nm

]

Electric Machine Operating Points

33

The fuel consumption of this architecture on NEDC cycle is 3.99 L/100km, which

represents a fuel saving of 25.8% compared to the conventional gasoline engine. The CO2

emissions are thus lowered to 96 gCO2/km. This challenges the Golf VII diesel, but is not

enough to reach the 2020 target.

3.5.3.2 P2 ARCHITECTURE

The P2 architecture is very similar to the P1 architecture excepted that an ICE side clutch

is added to make possible to physically decouple ICE and e-Machine. (Figure 3-12)

Figure 3-12 : P2 Hybrid Architecture.

Thus unlike the P1 architecture it is possible to run full electric or regenerate energy

during braking by disengaging the engine via the ICE side clutch. This architecture is

therefore more efficient than the previous one. Indeed as shown on Figure 3-13 the e-

machine is more used and even point at mid load are upshift to operate at better efficiency

(green box).

Figure 3-13 : Engine and Electric Machine Operating Points on NEDC (P2 Architecture).

The fuel consumption of this architecture on NEDC cycle is 3.52 L/100km, which

represents a fuel saving of 34.6% compared to the conventional gasoline engine. The CO2

1000 2000 3000 4000 5000 6000

0

50

100

150

200

w [RPM]

T [

Nm

]

Engine Operating Points

0 2000 4000 6000 8000 10000-80

-60

-40

-20

0

20

40

60

80

w [RPM]

T [

Nm

]

Electric Machine Operating Points

ICE Side Clutch

34

emissions are thus lowered to 86 gCO2/km. This architecture allows reaching the 2020

target. But the additional clutch required increases the space needed to implement this

topology. As the Golf VII has a transverse engine it is very difficult to make the side clutch

fit in the vehicle overhang, thus although this architecture presents high fuel benefits it is

not the most suitable for this project.

3.5.3.3 P3 & P4 ARCHITECTURE

The last architectures to be investigated are the P3 and P4 (Figure 3-14). These 2

topologies are studied together since they present the same pros and cons.

Figure 3-14 : P3 (left) and P4 (right) Hybrid Architecture.

As the motor is located after the manual gearbox, it is possible to perform regenerative

braking with engine totally decoupled. It is possible to set the E-Machine directly on the

transmission output shaft to save the additional weight induced by a double shaft, but this

will limit the torque at the wheels provided by the motor. But although a double shaft will

take more space and induce more weight it can fit with a transverse engine and can

enhance the motor torque. The advantage of the P3 architecture over the P4 architecture

is that the additional transmission can be downsized since the motor output torque will

be multiply by the final drive ratio.

To quantify the potential benefit of such topologies the P3 architecture will be studied

with a single shaft (ratio 1 with manual transmission output shaft) and with a double shaft

(ratio 3 with manual transmission output shaft).

Fuel Consumption (L/100km)

CO2 emissions

(g/km)

Fuel Savings (%)

P3 Ratio=1 4,04 96 25,0

P3 Ratio=3 3,89 93 27,7

Table 3-4 : P3 Architectures Fuel Consumption and Emissions.

35

As shown in Table 3-4 an increasing ratio between the motor shaft and the manual

transmission output shaft improves the fuel benefits thanks to a multiplication of the

motor torque at the wheels. As for the previous hybrid topologies the implementation of

an E-Machine allows avoiding the engine working at low load inefficient points (green box

on Figure 3-15).

Figure 3-15 : Engine and Electric Machine Operating Points on NEDC (P3 Ratio=3 Architecture).

As highlighted on Table 3-4 the fuel improvement realized with a P3 architecture and a

ratio of 3 is enough to meet the 2020 regulation.

3.6 FINAL ARCHITECTURE

Thanks to the previous investigation it is now possible to choose the most suitable

configuration for our project. The results of the previous study are summarized in Figure

3-16 and Figure 3-17.

Figure 3-16 : Hybrid Architecture Selection – Fuel Consumption.

1000 2000 3000 4000 5000 6000

0

50

100

150

200

w [RPM]

T [

Nm

]

Engine Operating Points

0 2000 4000 6000 8000 10000-80

-60

-40

-20

0

20

40

60

80

w [RPM]

T [

Nm

]

Electric Machine Operating Points

0

1

2

3

4

5

6

ConventionalGasoline

Stop & Start P1 P2 P3 R=1 P3 R=3

5,385,09

-5.7%3,99

-25.9% 3,52-34.6%

4,04-25.0%

3,89-27.7%

Fue

l Co

nsu

mp

tio

n (

L/1

00

km)

Hybrid Architectures

FUEL CONSUMPTION

36

Figure 3-17 : Hybrid Architecture Selection - CO2 Emissions.

These two figures clearly show that the best architecture in terms of fuel consumption

and emission is the P2 hybrid. But as previously stated this topology does not fit with a

transverse engine, which is the case for the Golf VII reference vehicle, due to the additional

clutch required. Thus since a double shaft is suitable for a transverse engine, the P3

architecture seems more legitimate to avoid a complete modification of the vehicle layout.

Moreover as shown on Figure 3-17 this architecture is able to reach the emission target

(95 gCO2/km) by playing with the additional transmission ratio.

The architecture being selected, it is now possible to optimize it by performing some

sensitivity studies on the E-Machine power and by implementing a new gear strategy in

order to further decrease the fuel consumption. As the vehicle is equipped with an E-

Machine able to provide additional torque to the wheel it is possible to choose the gear

shifting on NEDC and thus optimize it regarding fuel consumption.

0

20

40

60

80

100

120

140

ConventionalGasoline

Stop & Start P1 P2 P3 R=1 P3 R=3

128121

96

84

96 93

CO

2Em

issi

on

(g/

km)

Hybrid Architectures

CO2 EMISSIONS

2020 Regulation

37

4 GEAR SELECTION OPTIMIZATION

In this section is developed an optimized energy management strategy taking the gear

engaged as a second degree of freedom in addition to the engine torque. Moreover more

realistic assumptions will be considered: the Stop & Start function will be disabled during

the first 150 s of the cycle to allow the engine to warm up and the E-Machine will be used

to ensure the thermal engine cranking.

4.1 GEAR SHIFTING STRATEGY

As the transmission is manual, on real driving the gear shifting strategy optimized for fuel

consumption will only be seen as a gear shifting indicator (green box on Figure 4-1). The

driver can therefore choose if he desires to change gear or not.

Figure 4-1 : Gear Shifting Indicator.

4.1.1 PRINCIPLE

The gear shifting strategy implemented on Matlab/Simulink © is the following:

If the vehicle speed is constant or increase a gear has to be engaged. This condition

is required for safety, the vehicle cannot accelerate when the driver is not pressing

the acceleration pedal. This means that the E-Machine cannot run Full Electric with

thermal engine decoupled.

As the P3 architecture has been chosen, if the vehicle decelerates it is

recommended to shift to neutral to decouple the engine from the motor in order

to recover the maximum energy.

During take-off phase the E-Machine runs in Full Electric (1st gear) until the engine

launching speed is reached (750 RPM).

A fuel penalty is added to a gear shift to avoid oscillating gear shifting.

38

A fuel penalty is added to engine switch on/off to avoid oscillating behaviour.

The engine is not allowed to run below 1000RPM when the 2nd, 3rd, 4th or 5th gear

is engaged.

NB: In further developments it can be interesting to implement an e-clutch in order to be

able to disengage the ICE when running in full electric mode. In this way the engine

resistive torque will be deleted, without expecting the driver to press the clutch pedal.

4.1.2 IMPLEMENTATION

The previous Energy Management Strategy is used as a basis for this new strategy.

However instead of defining a vector of 12 candidate for the imposed gear, 6 vectors of 12

candidates are defined one vector for each of the gear. Then as in the previous

management strategy the solution that minimizes the Hamiltonian function will be chosen

and the gear related to this solution recommended.

Flags are raised to notify that a solution is not acceptable if it does not respect the

conditions stated in the section 4.1.1 .

4.2 SIMULATIONS

Simulations are performed on NEDC cycle and on a Real Driving Cycle.

4.2.1 NEDC CYCLE

The optimized gear shifting strategy is displayed on Figure 4-2. As expected, it is visible

on this graph that to save fuel the driver has to upshift as soon as possible.

39

Figure 4-2 : Gear Optimization on NEDC.

Figure 4-3 highlights the fuel flow required for the designed thermal engine on NEDC

Cycle. As previously stated during the first 150s the engine has an idle fuel consumption

due to the disabling of the Stop & Start function. The engine cranking phases are also

visible. Indeed at each vehicle take off there is a delay during which the E-Machine is

running full electric to allow the thermal engine reaching its launching speed.

Figure 4-3 : Fuel Flow on NEDC.

Regarding the battery state of charge, the final value is 0.58 which is close enough to 0.6

(initial SOC) to ensure the charge sustaining mode (Figure 4-4). During the first 150s the

battery is depleting constantly since there is no use of E-Machine during cranking phase

but electric power has to be supplied to auxiliaries (200W).

0

1

2

3

4

5

0 200 400 600 800 1000

Ge

ar (

-)

Time (s)

Gear Optimization on NEDC

Fix. Gears Opt. Gears

0,0E+00

5,0E-04

1,0E-03

1,5E-03

2,0E-03

2,5E-03

0

20

40

60

80

100

120

0 200 400 600 800 1000

Fuel

Flo

w (

kg/s

)

Veh

icle

Sp

eed

(km

/h)

Time (s)

Fuel Flow on NEDC

V_veh d_fuel

Engine cranking phase E-Machine runs Full Electric

40

Figure 4-4 : Battery State of Charge on NEDC.

The fuel consumption on NEDC for this new energy management strategy is 3.74

L/100km and the associated CO2 emissions are 89 gCO2/km. This represents a fuel

economy of 3.9 % compared to the same hybrid architecture (P3) but with the previous

energy management strategy gear shifting being imposed. This is due to the fact that this

new strategy allows engine operating points to be moved to better efficiency points

(green box on Figure 4-5).

Figure 4-5 : Engine and Electric Machine Operating Points on NEDC (P3 - Ratio=3 Architecture with

Optimal Gear Shifting).

This architecture emitting 89 gCO2/km on the NEDC cycle fulfill the requirement of 2020

regulation concerning pollutant emissions and is 10 gCO2/km better than the target state-

of-the-art Diesel Golf VII.

0,35

0,4

0,45

0,5

0,55

0,6

0,65

0

20

40

60

80

100

120

0 200 400 600 800 1000

SOC

(-)

Veh

icle

Sp

eed

(km

/h)

Time (s)

Battery State of Charge

V_veh SOC

1000 2000 3000 4000 5000 6000

0

50

100

150

200

w [RPM]

T [

Nm

]

Engine Operating Points

0 2000 4000 6000 8000 10000-80

-60

-40

-20

0

20

40

60

80

w [RPM]

T [

Nm

]

Electric Machine Operating Points

41

To continue the optimization process it could be interesting to perform a sensitivity study

on motor power since so far the power has been set to the maximum power of mild hybrid

vehicle (15 kW) and on transmission ratio between the motor shaft and manual

transmission output shaft.

These two parameters are actually linked since to ensure engine cranking with the motor

a sufficient torque has to be provided by the motor. Thus if motor power is reduced the

transmission ratio has to be increased to raise the torque multiplication factor between

the motor shaft and the wheels. And if the transmission ratio is reduced the motor power

has to be increased.

It turns out that to ensure engine cranking with E-Machine on NEDC cycle the

transmission ratio need to be at least 3 and the E-Machine Power at least 15 kW, so these

quantities can only be increased. As the E-Machine power cannot be further increased

(15kW max for Mild-Hybrid applications) the only lever to play with is the transmission

ratio. However if the transmission ratio increases too much at maximum vehicle speed

the E-Machine can be driven at too important rotational speed leading to failure. The

maximum rotational speed allowed by the motor is 18 000 RPM and the maximum vehicle

speed is around 190 km/h. Thus the upper limit for the transmission ratio is 3.3.

But to limit the size and the weight of the additional transmission, as a ratio of 3 is enough

to meet the project targets, it has been decided not to increase this ratio and keep a value

of 3.

Let’s now consider a real driving cycle to observe if the CO2 emissions are also acceptable

on day to day life rides.

4.2.2 REAL DRIVING CYCLE IN PARIS

The driving cycle chosen is displayed on Figure 4-6. The journey links IFP School to

Bobigny. It includes urban and extra-urban driving parts but the maximum speed is

limited to 90 km/h which is far below the maximum speed allowed on French motorways.

42

Figure 4-6 : Real Driving Cycle in Paris.

The vehicle speed has been recorded using a GPS watch. There is therefore some

uncertainties linked to the measuring device.

Looking at the cycles speed profile (Figure 4-7), it is clearly visible that the RDE cycle is

much more transient than NEDC cycle and this is especially why this cycle is criticized.

Figure 4-7 : NEDC & RDE Cycles Speed Profiles.

To estimate the benefits of the designed hybrid architecture with the optimized gear

shifting strategy, the fuel consumption of the conventional gasoline vehicle has been

assessed with a simple gear shifting strategy based on vehicle speed. The conventional

gasoline architecture has a fuel consumption of 4.60 L/100km on this cycle.

0

20

40

60

80

100

120

0 500 1000 1500 2000 2500 3000

Veh

icle

Sp

eed

(km

/h)

Time (s)

Cycles Speed Profiles

NEDC RDE

43

By simulating the final hybrid architecture along with the optimized gear shifting strategy

it is possible to reach 3.29 L/100km on this cycle. This represents a fuel saving of 28.5%

compared to the conventional gasoline vehicle. In the same way that on the NEDC cycle,

the E-Machine allows engine operating points to be moved to better efficiency points or

suppressed (Figure 4-8).

Figure 4-8 : Engine and Electric Machine Operating Points on NEDC Conventional Gasoline (Left) /

Proposed Hybrid Architecture (Right).

On this RDE cycle the proposed hybrid architecture emits 79 gCO2/km which is also

below the 2020 regulation limit (95 gCO2/km).

Thus at the end of this study it clearly appears that the proposed hybrid architecture is

able to meet the project target in terms of emission, that is to say having CO2 emissions

lower than 95 gCO2/km, on both NEDC and RDE cycles.

1000 2000 3000 4000 5000 6000

0

50

100

150

200

w [RPM]

T [

Nm

]

Engine Operating Points

1000 2000 3000 4000 5000 6000

0

50

100

150

200

w [RPM]

T [

Nm

]

Engine Operating Points

Hybridization

44

5 E-SUPERCHARGER IMPLEMENTATION

Through downsizing and turbocharging the conventional gasoline engine, hybridization

of the vehicle with an electric motor and optimization of the gear selection strategy, it is

possible to reduce the fuel consumption considerably in order to reach the CO2 emission

target that has been set for the project. However this improvement should not be achieved

in sacrifice of the driving performance of the vehicle. The fun-to-drive should be either

maintained or improved.

Turbocharger is a good enabler regarding improving the specific power of the engine. A

larger turbocharger leads to a greater specific power, while reducing the specific fuel

consumption. However, as the size of the turbocharger gets bigger, the response time of

the engine will be longer, then it will take more time for the driver to feel the torque he

wants when pressing the acceleration pedal. This means degradation of the driving

performance [9]. The potential of the electric supercharger technology to solve this

drawback has thus been explored in this project.

According to the research done by automotive companies such as Valeo, potential gains

through integrating an electric supercharger to the engine include, but are not limited to,

the following aspects:

• Allows extreme downsizing;

• Enables good transient response at any engine speed;

• Allows downspeeding without decreasing the maximum torque that could be

achieved;

• Enables low cost hybridization when coupled with the regenerative system.

A picture of an electric supercharger from Valeo is shown in Figure 5-1. The axle of the air

compressor is driven by a specifically designed electric motor, which is powered by a

battery pack on the vehicle.

Figure 5-1 : Electric-Assisted Supercharger from Valeo.

45

In this phase, simulations have been performed to predict the roll-on acceleration

performance of the vehicle, in the cases of implementing an e-Supercharger or not, and

also using an e-motor during the acceleration or not.

Due to limited data availability and the complexity of air-loop modeling, an e-

Supercharger of 4kW will be explored directly without further consideration on sizing.

5.1 SYSTEM LAYOUT AND REQUIREMENTS

The inlet temperature and pressure of an e-Supercharger is often limited, thus restricting

its location to the upstream of the turbocharger, as shown in Figure 5-2 [10]. The e-

Supercharger is not capable of handling full airflow at high engine loads/speeds, thus a

bypass route for the e-Supercharger might be required. Other layout or solutions also

exist, for example, a motor could be integrated with the turbocharger directly [11].

Figure 5-2: Layout of the 2-Stage Air Charging System with an E-Supercharger. [10]

As mentioned before, the e-Supercharger has to be powered by a battery pack on the

vehicle. Since an alternate current motor is often used, an AC/DC inverter is necessary to

achieve the current frequency inversion. While this AC/DC inverter is often integrated on

the e-Supercharger itself, a DC/DC converter might still be needed to achieve a voltage

step-up or –down from the electric network on the vehicle. A generator is needed to

generate the necessary energy which is then stored in the battery for consumption later.

In the present study, since a 48V battery pack has been adopted and a 48V e-Supercharger,

a DC/DC converter is not requested between them. On the other side, the generator’s role

could be played by the e-motor. A dedicated control module will be needed to control the

T

CA

C

C

eSC

Bypass

valve

Turbocharger

C

46

operation of the e-Supercharger and its cooperation with the turbocharger, but this

control issue will not be tackled in this project due to lack of time.

5.2 SIMULATION METHODOLOGY

Roll-on acceleration performance is a good indicator to represent the fun-to-drive of a

vehicle. The range of roll-on acceleration can be varied. Typical ranges on which different

vehicles are compared include 30-60km/h, 60-100km/h and 80-120km/h. Also, the gear

on which an acceleration is realized can be selected depending on the range. For these

study all the three ranges mentioned above have been investigated, while the 5th gear has

been adopted mainly but other possibilities have also been explored.

The torque requested during the roll-on acceleration could be supplied by both the

internal combustion engine and the electric motor. The logic of simulations for a

turbocharged engine and an engine with both a turbocharger and an e-Supercharger is

the same except for the need of using different data on turbo lags.

5.2.1 ENGINE TORQUE CALCULATION

During an acceleration, the torque supplied by the engine is limited by the maximum

torque at certain engine rotation speeds but also depends on the turbo lag. In this project,

the simulation is implemented through a step-by-step procedure, in which the torque in

each time step is recalculated according to the torque in the previous step, and the torque

increase during the time of a step. Figure 5-3 gives an illustration of the initial condition

of the roll-on acceleration and the different steps of the calculation.

Figure 5-3: Illustration of the Step-by-step Calculation Procedure.

47

Initial condition: N0 corresponds to the engine speed in the selected gear (e.g., the 5th

gear) at the initial speed (e.g., 80km/h). T0 is the torque able to run the vehicle at constant

speed of 80km/h, after which a roll-on acceleration is initiated. N1 is the engine speed

before which the engine torque is kept at a constant level T1, which is the torque that could

be reached instantaneously after the driver presses the acceleration pedal without any

delay. Since the torque request to run at constant speed is relatively small compared to

an acceleration starting from that speed, T1 is usually larger than T0.

Typical calculation step (duration of dτ = 0.05s): At the beginning of a step, the engine is

running at the speed Ni, giving the torque Ti.

Ni+1 at the end of this step can be obtained by recalculating the vehicle speed, based on:

The speed of the previous step;

The acceleration enabled by the torque of the previous step;

The gear engaged.

Ti+1 can be calculated based on Ti and the increase of torque during this step (of duration

dτ):

Ti+1 = Ti + dTi+1 (Equation 5-1)

Where dTi+1 = dτ * T’(Ni+1, T i).

T’(Ni+1, T i) is the slope at torque Ti on the Time to Torque curve at speed Ni+1. However,

the evolution of the Time to Torque Curve has to be set first.

Figure 5-4: Typical Shape of the Time to Torque Curve at a Constant Engine Speed. [12]

48

Modeling of the Time to Torque curve: The typical shape of the Time to Torque curve

obtained by experiments on engine test benches is shown in Figure 5-4 [12]. It can be

divided into two phases. The first phase is the constant torque phase, which is quite short

and results from the lag of the air loop. In this phase the engine torque is instantaneously

increased to a certain level and then kept for a duration tt0. The second phase reflects the

ramp of the engine torque up to the maximum torque that could be reached at this

particular engine speed. As the torque gradient increase in the beginning of this phase is

usually large and tends to decrease with time until reaching zero, the exponential function

might be a good choice to approximately model this behavior. An exponential function has

thus been adopted for this study.

The complete function implemented to model the Time to Torque curve is given below:

0

0 0

0

0

, *

,

,* * *( )

i

i

max i max i i

T NT t N

a T N a T N

t tt

T N exp c t ttt tt

(Equation 5-2)

Where a and c are constants that indirectly set the time needed to reach full load tt.

This seems to be an explicit function which can approximately match the trend of the

actual turbo lag phenomenon. To keep it simple, it can be assumed that the overshoot

parameter a is set to 1. Then only c needs to be determined.

Figure 5-5: Curves Obtained by Changing Parameter c in the Time to Torque Function (a=1, tt0=0).

In practice, the data known are the initial torque 0 iT N , the duration of the constant

torque period, the maximum torque and the time needed to reach 90% maximum torque

at fixed certain engine rotation speeds. Based on this point of 90% maximum torque, the

constant c can be estimated for each engine speed:

49

0 00.9 * * ( ) ( ) ( ( ) ( )) * *( )max i max i max i iT N a T N a T N T N exp c tt tt

max max max 0 0ln(( * 0.9 ) / * ( )( ) ( ) ( ( ) ( ))) /i i i ic a T N T N a T N T N tt tt

Once c is known, the Time to Torque curve gradient at any point can be calculated. This

gradient is used in the step-by-step methodology previously described to calculate the

torque increase at each step.

5.2.2 MOTOR TORQUE

The torque coming from the e-motor can be obtained by interpolation using the motor

map, knowing the motor speed. As the motor is powered by the battery pack, attention

needs to be paid at each step of calculation to the State of Charge level of the battery. It

should always be kept above the specified lower limit to prevent any premature aging.

Figure 5-6: Interpolation on the Motor Characteristic Curve to Obtain the Motor Torque.

5.3 PERFORMANCE CALCULATION RESULTS

By implementing the methodology described in the previous section, the roll-on

acceleration performance on different ranges is obtained. Comparisons will be focused on

the difference between implementing or not an e-Supercharger, and also the difference

between implementing or not an e-motor.

5.3.1 TIME TO TORQUE AT CONSTANT ENGINE SPEEDS

The value of tt for each engine speed can be interpolated from the Time to Torque data at

several engine speeds. Interpolations have been done respectively for the turbocharger

and a 4kW e-Supercharger from Valeo, as shown in Figure 5-7.

50

Figure 5-7: Time to torque vs engine speed.

Using the formulas proposed in the previous section, the torque and thus the BMEP

evolution paths at 1250rpm and 3000rpm are plotted respectively in Figure 5-8 and

Figure 5-9. It can be seen that at 1250rpm, there is a big gap between the implementation

with and without an e-Supercharger, while at 3000rpm the two torque evolution paths

are actually quite close. This indicates that there is a negligible impact of the e-

Supercharger when the engine is running at high rpm.

Figure 5-8: Torque Evolution at 1250rpm.

Figure 5-9: Torque Evolution at 3000rpm.

0,0

0,5

1,0

1,5

2,0

2,5

3,0

3,5

4,0

4,5

1000 1500 2000 2500 3000 3500

Tim

e to

max

imu

m t

orq

ue

(s)

Engine speed (rpm)

Response time without eSC

Response time with eSC

0

5

10

15

20

25

0,0 1,0 2,0 3,0 4,0 5,0 6,0 7,0 8,0 9,0

BM

EP (

bar

)

t (s)

TC+Esc

TC

0

5

10

15

20

25

0,0 1,0 2,0 3,0 4,0 5,0 6,0 7,0 8,0 9,0

BM

EP (

bar

)

t (s)

TC+Esc

TC

51

5.3.2 OVERALL RESULTS ON ALL ACCELERATION RANGES CONSIDERED

Figure 5-10 gives a summary of all calculation results for the different ranges of roll-on

acceleration, i.e. from 80-120km/h, 60-100km/h and 30-60km/h. All these accelerations

are performed on the 5th gear with a reduction ratio of 3.05 from the engine to the wheels

(V1000=38km/h/krpm).

Figure 5-10 : Roll-on Performance of the Vehicle with Different Architectures.

Roll-on 80-120km/h: When the e-motor is not available or turned off during the

acceleration, a 4% improvement is observed for the architecture with an e-Supercharger

compared with the solution with only a turbocharger. If the e-motor is turned on, the time

needed to reach 120km/h could be shorten considerably, while an e-Supercharger can

still achieve around 4% improvement additionally.

Roll-on 60-100km/h: The e-Supercharger helps saving around one second without any e-

motor assistance, and can give a reduction delay of 6.5% when the e-motor is used.

Roll-on 30-60km/h: The gain is maximum for this test. When no torque is asked from the

e-motor, 1.35 seconds could be saved by implementing the 4kW e-Supercharger, which

means around 10% improvement.

To summarize, the Electric-assisted supercharger is able to considerably reduce the time

to torque, and this improvement is especially interesting when the engine is running at

low engine speed. On the other side, the e-motor is another good enabler of better

acceleration performance, however as we will see later, it requires rapid depletion of the

battery capacity.

14.0113.42

14.0513.46

12.42 12.70

9.82

8.88

5.97

9.46

8.30

5.52

5

6

7

8

9

10

11

12

13

14

15

80-120km/h 60-100km/h 30-60km/h

Du

rati

on

(s)

Ranges Of Roll-on Acceleration

TC w/o EM

eSC w/o EM

TC w/i EM

eSC w/i EM

Diesel Golf VII

52

To conclude implementing a 4kW e-Supercharger alone is not able to gain enough

advantage over the Golf VII 1.6 TDI model, the 15kW e-motor is needed additionally. This

might be attributed to the short gears we have selected and to the fact that the interaction

between the e-Supercharger and the rest of the air loop components are not taken into

account.

5.3.3 FOCUS ON 80-120KM/H ACCELERATION

In this section detailed information will be given for roll-on acceleration from 80 to

120km/h, to enable some further analysis and discussion.

Figure 5-11 gives the engine torque evolution paths during the acceleration from 80 to

120km/h, respectively for the architecture with and without an e-Supercharger. Notice

that the concerned range of engine speed is between 2000rpm and 3200rpm. Recalling

Figure 5-7, this means the difference in time to reach maximum torque is between 0.25s

and 1.2s. Thus the final difference in the performance is limited to 1.2s, and should be in

fact lower. This explains why we did not gain so much with an e-Supercharger (Figure

5-10).

Figure 5-11: Engine Operating Point Evolution (without Motor).

It is also possible to plot the vehicle acceleration evolution for different cases (Figure

5-12). This gives a more clear view on where the differences come from. Velocity

evolution profiles are plotted in Figure 5-13 to illustrate the roll-on experiences for

various cases.

50,0

70,0

90,0

110,0

130,0

150,0

170,0

190,0

210,0

1000 1500 2000 2500 3000 3500

Torq

ue

(Nm

)

Engine speed (rpm)

TC+eSC

TC

Max Torque [Nm]

53

Figure 5-12: Vehicle Acceleration Evolution.

Figure 5-13: Vehicle Speed Evolution.

Let’s now consider the state of charge (SOC) of the battery pack during the accelerations.

As highlighted on Figure 5-14, the main electricity consumption comes from the operation

of the e-motor. When the e-motor is on, the SOC of the battery can decrease by 10% in less

than 10 seconds. This might not be good for needs to run in pure-electric mode later to

enable better fuel economy. In contrast, the e-Supercharger leads to only 3% of SOC

decrease when the target speed is reached. This is consistent with the sized power levels

of the e-motor (15kW) and the e-Supercharger (4kW).

0,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

Acc

eler

atio

n (

m/s

)

t (s)

TC w/o motor

eSC w/o motor

TC w/i motor

eSC w/i motor

75,0

80,0

85,0

90,0

95,0

100,0

105,0

110,0

115,0

120,0

125,0

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

velo

city

(m

/s)

t (s)

TC w/o motor

eSC w/o motor

TC w/i motor

eSC w/i motor

54

Figure 5-14: Battery SOC Evolution.

Finally, the potential impact of changing the gear ratio has been investigated. In Figure

5-10 it can be seen that with the e-motor, the roll-on performances are actually far better

than the state-of-art diesel engine. This means the gear ratios can be further increased in

order to reach further fuel-economy benefit, while keeping a good fun-to-drive.

Let’s now change the 5th gear ratio (V1000) from 38 to 47.5km/h/krpm. Table 5-1

highlights that after lengthening the final gear ratio, the performance of the architecture

with an e-Supercharger can still reach the target derived from the competitor, while

implementing a turbocharger only is no longer able to meet the target. The benefit

brought by the e-Supercharger is enlarged now compared with the shorter gear ratio,

since the engine is now running at lower speeds. This gear ratio increase is not in vain

since it could bring even more fuel consumption gain, which is the first-priority concern.

V1000 5th gear (km/h/krpm) 38 47.5

Architecture Turbo only Turbo + eSC Turbo only Turbo + eSC

Time needed from 80-120km/h (s) 9.82 9.46 12.16 11.54

Table 5-1: Effects of Varying the Gear Ratio (E-Motor Active).

Thus if more time was allocated to the project it could be interesting to perform an

optimization on the manual transmission (gear length, addition of a 6th gear).

40,0%

45,0%

50,0%

55,0%

60,0%

65,0%

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

SOC

t (s)

TC w/o motor

eSC w/o motor

TC w/i motor

eSC w/i motor

55

6 RESULTS

The proposed gasoline hybrid needs to fulfill emission standards, cost to the OEM, and

satisfy the customer expectations in terms of performance and cost better than the

existing Diesel Golf VII. The comparison in each domain is individually discussed.

6.1 FUEL CONSUMPTION & CO2 EMISSIONS

6.1.1 NEDC CYCLE

Figure 6-1 : NEDC Cycle Fuel Consumption & Emissions.

From the above plot we can observe a 32.5% decrease in CO2 emissions compared to the

conventional gasoline and a 10 % decrease in emission compared to the baseline Diesel

Golf VII. Hence from a legislative point of view a clear benefit is observed and is good in

terms of marketing strategies.

6.1.2 RDE CYCLE

The RDE cycle is implemented in the model and the emission values obtained show a 30%

reduction between the conventional gasoline and the final proposed hybrid architecture.

5,5

3,8 3,74

132

9989

0

20

40

60

80

100

120

140

0

1

2

3

4

5

6

Conventional Gasoline Diesel Proposed Hybrid

CO

2Em

issi

on

s (g

/km

)

Fue

l Co

nsu

mp

tio

n (

L/1

00

km)

NEDC CYCLE

Fuel Consumption ( L/100km) C02 emission

10%

56

Figure 6-2 : RDE Cycle Fuel Consumption & Emissions.

6.2 SYSTEM OVERVIEW

Figure 6-3 : System Overview.

The system is a P3 hybrid architecture implemented in a transverse frontal gasoline

engine layout. The system consists of a 5 speed manual transmission and an e-motor is

linked to the transmission through a gear (in our case ratio: 3). There are two battery

packs (12V and 48 V) to the rear of the car and the circuit is built using a DC/DC inverter.

The other conventional layout remains the same. An e-Supercharger is linked to the

engine in the front and is powered by the e-machine.

Tank

48V Battery

DC/DC

Inverter

12V Battery

12V Electric

Loads

E-Supercharger

(4kW)

1.2L 3-cylinder

Gasoline Engine

Electric Motor

(15kW)

5-speed Manual

Gearbox

57

With respect to the conventional gasoline vehicle the main additional components are:

• E-Supercharger (4kW)

• E-motor (15 kW)

• Battery pack (48V)

• DC/DC converter

• Cables

The weight and cost addition due to the above mentioned are discussed in the next

section.

6.3 ADDITIONAL WEIGHT & COST ESTIMATION

The weight of individual components described above and a rough indicated market price

are listed. The price is and indicative market price for 2020 for mass production (> 200

000 pieces/year). These give a total of a 43 kg weight addition and a 1285 € cost addition.

However these are just approximate values and may differ with OEM and supplier.

6.4 COST TO THE OEM

The CO2 reduction cost is an important value to the OEM and the lower CO2 emissions

cost justifies the choice of implementing a gasoline hybrid version for a Diesel car. The

assumptions made in this synthesis include that the Diesel engine is 40 kg heavier than a

gasoline version 3-cylinder and cost about 1000 euros more.

Comparison: Proposed

Hybrid

Conventional

Gasoline

vs. Diesel (Golf VI 1.6TSI)

Weight Increase + 43 kg + 3 kg

Additional Cost + 1285 € + 285 €

CO2 Benefit (g/km) 43 g/km 10 g/km

CO2 Reduction Cost 32.1 €/gCO2/km 28.5 €/gCO2/km

Table 6-1 : Cost to the OEM.

58

6.5 TOTAL COST OF OWNERSHIP ESTIMATION

In Table 6-2 are listed the assumptions used to estimate the total cost of ownership.

Gas Hybrid Diesel

Mean distance in France (km/year)

12 700 12 700

Ownership duration (year)

5 5

Fuel rate (€/L) (Super 98) 1.28 1.17

Fuel consumption (L/100km)

3.72 3.9

Table 6-2 : Total Cost of Ownership Assumptions.

By statistical surveys the mean ownership duration in Europe for a car is found to be a

period of 5 years. And the present day average fuel rate throughout Europe for gasoline

is 1.28 €/L and diesel is 1.17 €/L. From these values the total cost spent by the customer

on fuel can be estimated. The values of gasoline and diesel are averaged and are

considered for the year 2015.

Cost Calculation Gas Hybrid Diesel

Insurance (€/year) 385 550

Maintenance (€/year) 900 1000

Fuel consumption (€/year) 605 560

Total cost (€/5years) 9450 10550

Table 6-3 : Total Cost of Ownership Calculation.

The rough average amount a customer will tend to spend on the vehicle throughout its

lifetime is found to be lesser for the proposed final hybrid architecture with a net saving

of 1100 €/ 5 years. This data is based on the fact that market in Europe is Hybrid friendly

meaning the insurance is lesser for hybrid cars. The maintenance is assumed lesser than

Diesel because of better engine operating conditions and lesser soot.

59

CONCLUSION AND DIRECTIONS FOR FURTHER STUDY

Through this project, mild hybridization and e-Supercharger technology have been

applied to a mid-size gasoline powered vehicle in order to reach both the emissions and

driving performance level of a state-of-art diesel vehicle model (Golf VII 1.6 TDI).

To reach these targets a 1.6L 4-cylinder gasoline engine has first been downsized to

decrease the engine displacement while keeping the output power necessary to reach

good performance level. Then the fuel consumption gain of decreasing the number of

cylinders to three has been explored by using an explicit heat losses model. The

advantages and disadvantages of various hybrid architectures have been compared, and

main efforts have been deployed on the sub-categories of the parallel hybrid. A Simulink

© model has been built to compare the fuel consumption gain related to each hybrid

topology. At the end of this study a non-coaxial P3 architecture has been selected, since it

features both good fuel consumption gain and proper compatibility with the transverse

layout of the C-segment cars. At the same time, sizing of the electric motor and battery has

been done to satisfy various constraints. Optimization of gear selection and energy

management strategy has been implemented, in order to achieve better fuel economy on

the traditional NEDC cycle and a Real Driving Cycle established by the team. To keep or

even improve the fun-to-drive, the potential of e-Supercharger technology has been

investigated. Explicit models have been built to model the time to torque behavior and a

step-by-step methodology has been implemented to simulate the roll-on performance of

the proposed hybrid on different ranges of acceleration. Finally, a cost estimation has

been done to evaluate the balance between gain and cost.

Main findings and conclusions are as follows:

• Through Mild Hybridization (48V) and optimization of the energy management

strategy, it is possible to achieve good CO2 emissions (89 gCO2/km), better than both

conventional gasoline and state-of-the-art diesel vehicle and thus meet the 95

gCO2/km target.

• By implementing both an electric machine (15kW) and an electric supercharger

(4kW), the driving performance of the proposed hybrid can be maintained or even

improved.

60

• Although the price of a mild hybrid can be more expensive compared with a

conventional diesel, benefits can be gained back during its lifetime.

Thus this study indicates that the solution of 48V hybridization of a mid-size gasoline

vehicle using an electric-assisted supercharger is promising, and poses good

competitiveness over a state-of-art diesel.

Due to the time limit, the optimization process has not been tackled in an exhaustive way.

Further works remain to be done to make the results more accurate and reliable. This

includes:

• A detailed modeling of the air-loop to predict more accurately the performance of the

combination of a turbocharger and an e-Supercharger. This will allow to take into

account the interaction between the e-Supercharger and the turbocharger and to

simulate the BSFC improvement due to engine back pressure reduction when using an

e-Supercharger.

• An iteration process to further optimize the gear ratios and find the best trade off in

terms of fuel consumption and fun to drive.

• Fuel benefit estimation for the proposed hybrid architecture on other cycles such as

WLTC.

61

REFERENCES

[1] Heywood, John, B. (1988) “Internal Combustion Engine Fundamentals”, McGRAW-HILL

INTERNATIONAL EDITIONS.

[2] Millo, F., Badami, M., Ferraro, C.V., Lavarino, G. and Rolando, L. (2010), “A Comparison

Between Different Hybrids”, SAE Technical Papers.

[3] Rajendran Vallur, A., Khairate, Y. and Awate, C. (2015), “Prescriptive Modeling,

Simulation and Performance Analysis of Mild Hybrid Vehicle and Component Optimization”,

SAE Technical Papers.

[4] Nesscap (2015), “Ultracapacitor Data Sheet”, available at:

http://www.nesscap.com/ultracapacitor/EDLC/Supercapacitor/Large_cell_supercapaci

tor_family/cylindrical_supercapacitor_cell.jsp [accessed on 30/05/15]

[5] Kasseris, E. P. and Heywood, J. B. (2007), “Comparative Analysis of Automotive Powertrain”, SAE Technical Papers. [6] Sciarretta, A. (2015), “Hybrid Energy Management relevant to Powertrain Control and Mechatronics (TU10)”, IFP School: unpublished. [7] Levassor, W., Wysoki, L. (2015), “Longitudinal Dynamics Practical Works relevant to Engine Technology (TU4)”, IFP School: unpublished. [8] Ravello, V. (2015), “Hybrid Drive Train relevant to Transmissions and Alternative Drive Trains (TU8)”, IFP School: unpublished.

[9] Yamashita, Y., Ibaraki, S., Sumida, K. (2010), “Development of Electric Supercharger to Facilitate the Downsizing of Automobile Engines”, Mitsubishi Heavy Industries Technical Review, Vol. 47 No. 4.

[10] Speaker from MAHLE Powertrain (2015), “eSupercharging for Heavily Downsized Gasoline Engines”, SIA Conference Report.

[11] Prasad, S. D., Beshah, A. and Robert, P. (2010), “Coordinated Electric Supercharging and Turbo-Generation for a Diesel Engine”, SAE Technical Papers.

[12] Petitjean, D. (2015), “Turbocharger technology and intake supercharging for engines

(TU5)”, IFP School: unpublished.