Towards Low-Voltage, High-Current: A pioneering drive concept for battery electric vehicles

(1)

©2021 by

Stefan Haller

Towards Low-Voltage, High-Current

A pioneering drive concept for battery electric vehicles

Stefan Haller

Main supervisor: Prof. Kent Bertilsson Co-supervisor: Prof. Bengt Oelmann

Dr. Peng Cheng

Faculty of Science, Technology and Media Department of Electronics Design

Thesis for the Degree of Doctor of Philosophy Mid Sweden University, Sundsvall, Sweden, 2020

(2)

©2021 by

Stefan Haller

teknologie doktorsexamen i Elektronik måndagen den 11 januari 2021, klockan 8.30 i sal C306, Mittuniversitetet Sundsvall.

Seminariet kommer att hållas på engelska.

Towards Low-Voltage, High-Current

A pioneering drive concept for battery electric vehicles

Printed by Mid Sweden University, Sundsvall December 15, 2020

ISSN 1652-893X

ISBN 978-91-88947-85-7

Faculty of Science, Technology and Media

Mid Sweden University, Holmgatan 10, SE-851 70 Sundsvall, Sweden Phone: +46 (0)10 142 80 00

Mid Sweden University Doctoral Thesis 337

(3)

©2021 by

Stefan Haller

“The most dangerous phrase in the language is

‘We’ve always done it this way.’ ” – Grace Hopper

(4)

©2021 by

Stefan Haller

This thesis has been typeset using L^ATEX.

(5)

©2021 by

Stefan Haller

ABSTRACT

The first electric low-voltage vehicles were constructed in the mid-19^th century, but by the early 20^thcentury they were progressively replaced by successors with internal combustion engines. As the consequences of using fossil fuels are better understood, our society is now transitioning back.

The strong driving force towards electric transportation can be traced to several events and trends. The foremost of these is perhaps the rising awareness of climate change and the necessary reduction of the environmental footprint, as well associated political will for change.

Alongside this, the pioneering automotive company Tesla, Inc. showed what electric cars are capable of and how to easily charge them along the road. The diesel gate unearthed in 2015, also played a major role.

This transition is not without challenges, however. An electric car is expected to be reasonable priced, sustainable, environmentally friendly and electrically safe, even in case of an accident. Overnight charging at home should be possible, as well as the ability to quickly charge while in transit.

While the industry has long experience with high-voltage electrical machines, the required battery technology is quite new and low-voltage in nature. Currently, the battery is the most costly part of an electric drivetrain and it has the highest environmental impact. Efficient battery use is therefore key for sustainability and a responsible consumption of the resources available.

Nonetheless, most electric vehicles today use lethal high-voltage traction drives which require a considerable isolation effort and complex battery pack. Previous research results showed that a 48 V drivetrain compared to a high-voltage one, increases the drive-cycle efficiency.

Hence, similar driving range can be reached with a smaller battery.

This thesis provides an introduction to low-voltage, high-current, battery-powered traction drives. With the aim of increasing efficiency, safety and redundancy while reducing cost, a solution that breaks with century-old electric machine design principles is proposed and investigated. An overview and motivation to further investigate 48 V drivetrains with intrinsically safe and redundant machines is provided.

The main focus of this work is the practical implementation of multi- phase low-voltage but high-current machines with integrated power

(6)

©2021 by

Stefan Haller

electronics as well as components for a 48 V drivetrain. With this work, it is confirmed that today’s MOSFETs are not the limiting factor towards low-voltage, high-current drives.

In the first part of this work, two small-scale prototype machines were constructed and tested. The air-cooled, small-scale 1.2 kW prototype reached a copper fill-factor of 0.84. The machine’s low terminal- to-terminal resistance of 0.23 mΩ, including the MOSFET-based power electronics, allowed continuous driving currents up to 600 A. The resistive MOSFET losses stayed below 21 W.

The second part focuses on the key components for a 48 V high- power drivetrain. A W-shaped coil for a multiphase 48 V machine with direct in-conductor cooling was designed and tested. With glycol water, it reached a current density of 49.5 A/mm²with 0.312 l/min flow rate. Furthermore, a reconfigurable battery pack for 48 V driving and high-voltage charging was investigated.

(7)

©2021 by

Stefan Haller

SAMMANFATTNING

De första elektriska lågspänningsbilarna konstruerades i mitten av 1800-talet men ersattes succesivt med förbränningsmotorer i början av 1900-talet. Nu när konsekvenserna av storskalig användning av fossila bränslen börjar synas blir elektriska fordon återigen aktuella.

Den starka drivkraften mot elektrifierad transport kan kopplas till ett flertal händelser och trender. Främst är troligen den ökade medve- tenheten kring klimatförändringar och den nödvändiga minskningen av samhällets klimatavtryck, tillsammans med den politiska föränd- ringsviljan. Samtidigt visade den banbrytande biltillverkaren Tesla, Inc.

vilken prestanda elektriska bilar kunde förmå samt hur lätt de kunde laddas ute på vägarna. Dieselskandalen som uppdagades 2015 bidrog också på ett betydande sätt.

Övergången sker dock inte utan utmaningar och en elbil förväntas att vara prisvärd, hållbar, miljövänlig och elektriskt säker, även vid en olycka. Den ska kunna laddas över natten hemma samt snabbt utefter vägen.

Medan industrin har lång erfarenhet av elektriska motorer som arbetar med hög spänning så är den erforderliga batteritekniken ganska ny och lågspänd till sin natur. För närvarande är batteriet den dyraste delen av en elektrisk drivlina och dessutom har den en stor negativ miljöpåverkan. För ett hållbart och ansvarsfullt utnyttjande av våra resurser måste batteriet användas så optimalt som möjligt.

Trots detta, använder de flesta elfordon idag en dödligt hög spänning för sin drivlina vilket ger en komplex konstruktion av batteripaketet och många säkerhetsaspekter måste uppfyllas. Tidigare forskning visar dock att en 48 V drivlina kan uppnå högre effektivitet jämfört med en högspänd drivlina.

Denna avhandling ger en introduktion till lågspända elektriska drivlinor för elektrifierade fordon. Målet är att öka verkningsgrad, säker- het, tillförlitlighet och samtidigt reducera kostnader, vilket är en lösning som bryter mot de traditionella konstruktioner som använts i över ett århundrade. Avhandlingen syftar till att ge en översikt och motivation för att ytterligare undersöka denna typ av elektriskt ofarliga och redun- danta motor för elfordon. Arbetet fokuserar på en praktisk konstruktion av en mångfasig elektrisk motor med integrerad kraftelektronik för höga strömmar. Avhandlingen visar att dagens MOSFETs inte är den

(8)

©2021 by

Stefan Haller

begränsande faktorn i en lågspänd drivlina med höga strömmar.

I den första delen av detta arbete har två småskaliga prototypma- skiner byggts och testats. Den luftkylda småskaliga 1,2 kW-prototypen hade en kopparfyllningsfaktor på 0,84 och det låga terminal-till-terminal motståndet på 0,23 mΩ, inklusive den MOSFET-baserade kraftelektro- niken, tillät en kontinuerlig driftström på upp till 600 A. Dom resistiva MOSFET-förlusterna var mindre än 21 W.

Den andra delen fokuserar på nödvändiga komponenter som krävs till en 48V drivlina för höga effekter. En W-formad direktkyld ro- torlindning, för en flerfasig 48V-maskin, har designats och testats. Med flödande etylenglykolvatten genom lindningen uppnåddes en ström- täthet av 49,5 A/mm² vid 0,312 l/min flödeshastighet. Utöver detta utforskades ett omkonfigurerbart batteripaket för 48 V vid körning och hög spänning vid laddning.

(9)

©2021 by

Stefan Haller

©2021 by

Stefan Haller

2.2 Problem Formulation . . . 21

2.2.1 LVHC Machines with Novel MOSFETs . . . . 22

2.2.2 Improvement and Limits of LVHC Machines 23 2.2.3 Components for 48 V LVHC Machines . . . . 23

2.3 Objective . . . 23

2.3.1 Limitations . . . 24

3 Publications Included in this Thesis 27 3.1 Air-Gap Flux Density Measurement System for Verifi- cation of PM Motor FEM Model . . . 28

3.2 Investigation of a 2 V 1.1 kW MOSFET Commutated DC Motor . . . 29

3.3 Initial Characterization of a 2 V 1.1 kW MOSFET Com- mutated DC Motor . . . 29

3.4 A 2.5 V 600 A MOSFET-Based DC Traction Motor . . 30

3.5 Phase Current Shift in Multiphase Single-Turn Con- centrated Windings . . . 31

3.6 CPLD and dsPIC Hybrid-Controller for Converter Prototyping . . . 33

3.7 Multi-Phase Winding with In-Conductor Direct Cool- ing Capability for a 48 V Traction Drive Design . . . . 33

3.8 Reconf. Battery for Charging 48 V EVs in High-Voltage Infrastructure . . . 35

4 Theory 37 4.1 Introduction to Electrical Machines . . . 37

4.1.1 Synchronous Machines . . . 38

4.1.2 Asynchronous Machines . . . 39

4.2 EMF Generation . . . 40

4.3 Torque Production . . . 41

4.3.1 Magnetic and Reluctance Torque . . . 42

5 Methods 43 5.1 Paper I to Paper III, Prototype Machine 1 . . . 43

5.2 Paper VI and Paper V, Prototype Machine 2 . . . 44

5.3 Paper VI and Paper VII, Next Generation 48 V . . . . 46

6 Discussion 47

(11)

©2021 by

Stefan Haller

Contents

6.1 Machine Prototypes . . . 48

6.1.1 1^stPrototype . . . 49

6.1.2 2^ndPrototype . . . 50

6.2 Integrated Power Electronics . . . 51

6.3 Armature Reaction causing Current Unbalance . . . 53

6.4 Direct Winding Cooling . . . 54

6.5 Next Generation 48 V Drive . . . 54

6.6 Social and Ethical Considerations . . . 55

7 Outlook and Conclusion 57 7.1 Conclusion . . . 57

7.2 Future Work . . . 60

Acronyms 61 Bibliography 63 II Included Papers 73 Paper I — Air-Gap Flux Density Measurement System for Verification of PM Motor FEM Model 75 i.1 Introduction . . . 77

i.2 Existing Measurement Methods . . . 79

i.3 Design and Implementation . . . 80

i.3.1 Sensor Board . . . 80

i.3.2 Amplifier Board . . . 81

i.4 Experimental Setup . . . 81

i.4.1 Transducer Normalization . . . 84

i.5 Measurement Results . . . 84

i.6 Conclusion . . . 85

i.6.1 Future Work . . . 88

i.7 References . . . 89

Paper II — Investigation of a 2 V 1.1 kW MOSFET Commutated DC Motor 91 ii.1 Introduction . . . 93

ii.2 Proposed Architecture . . . 95

(12)

©2021 by

Stefan Haller

ii.3 Prototype Design . . . 96

ii.3.1 Electronic Commutator Design . . . 96

ii.3.2 Outrunner Rotor Design . . . 99

ii.3.3 Stator Design . . . 99

ii.4 Experimental Setup . . . 101

ii.5 Measurement Procedure and Results . . . 102

ii.5.1 Stator Coil Resistance . . . 102

ii.5.2 Locked Rotor Torque Measurement . . . 103

ii.5.3 Motor Constants kE and kT . . . 104

ii.5.4 Determine Motor Constant kT . . . 106

ii.5.5 Determine Motor Constant kE . . . 106

ii.5.6 No-Load Current . . . 108

ii.5.7 No-Load Power Consumption . . . 109

ii.6 Conclusion and Future Work . . . 109

ii.7 References . . . 110

Paper III — Initial Characterization of a 2 V 1.1 kW MOSFET Commutated DC Motor 113 iii.1 Introduction . . . 115

iii.2 Proposed Drive System . . . 117

iii.3 Experimental Setup . . . 119

iii.4 Measurement Procedure and Results . . . 120

iii.4.1 Stator Coil Resistance Measurement . . . 120

iii.4.2 Locked Rotor Torque Measurement . . . 121

iii.4.3 Steady State Efficiency Measurement . . . 122

iii.4.4 Efficiency and Loss Map . . . 123

iii.5 Conclusion . . . 126

iii.6 References . . . 127

Paper IV — A 2.5 V 600 A MOSFET-Based DC Traction Motor 129 iv.1 Introduction . . . 131

iv.2 Prototype Design . . . 133

iv.3 Experimental Setup . . . 136

iv.4 Measurement Procedure and Results . . . 138

iv.4.1 Stator Phase Resistance Measurement . . . . 138

iv.4.2 Efficiency Measurement . . . 140

(13)

©2021 by

Stefan Haller

Contents

iv.4.3 Torque-Speed Efficiency Map . . . 140

iv.4.4 Loss Distribution Map . . . 140

iv.5 Conclusion . . . 144

iv.6 References . . . 145

Paper V — Rebalancing of Phase Current Shift Caused by Armature Reaction in Multiphase Single-Turn Con- centrated Winding Machines 147 v.1 Introduction . . . 149

v.1.1 Resistive Losses and Fill-Factor . . . 150

v.1.2 Armature Reaction on Single-Turn Windings 150 v.1.3 Focus of this Study . . . 151

v.2 LVHC Machine . . . 151

v.2.1 Machine Design . . . 152

v.2.2 Winding Design . . . 152

v.2.3 Power Electronics . . . 152

v.2.4 Commutation . . . 155

v.2.5 Commutation Strategy . . . 157

v.3 Simulation . . . 158

v.4 Experimental Setup . . . 159

v.5 Results and Discussion . . . 162

v.5.1 Measurements . . . 163

v.5.2 Resistive and Switching Losses . . . 163

v.5.3 LVHC Machine Improvements . . . 166

v.6 Conclusion and Future Work . . . 168

v.6.1 Conclusion . . . 168

v.6.2 Future Work . . . 169

v.7 References . . . 169

Paper VI — CPLD and dsPIC Hybrid-Controller for Con- verter Prototyping 173 vi.1 Introduction . . . 175

vi.2 Proposed Controller . . . 177

vi.2.1 Signal Delay . . . 178

vi.3 Experimental Setup . . . 179

vi.3.1 Configurable Transformer PSFB . . . 179

vi.4 Experimental Results . . . 180

(14)

©2021 by

Stefan Haller

vi.5 Simulation Results . . . 185

vi.6 Conclusion . . . 187

vi.7 References . . . 187

Paper VII — Multi-Phase Winding with In-Conductor Direct Cooling Capability for a 48 V Traction Drive Design 189 vii.1 Introduction . . . 191

vii.2 Winding Design . . . 192

vii.3 Analytical Calculation . . . 194

vii.4 FEM/CFD Simulation Setup . . . 197

vii.5 Measurement Setup . . . 198

vii.6 Results . . . 201

vii.7 Conclusion . . . 205

vii.8 References . . . 206

vii.9 Biographies . . . 207

Paper VIII — Reconfigurable Battery for Charging 48 V EVs in High-Voltage Infrastructure 209 viii.1 Introduction . . . 211

viii.2 Battery Reconfiguration . . . 214

viii.2.1 Prototype Design . . . 214

viii.2.2 Battery Pack Balancing . . . 218

viii.3 Experimental Results . . . 218

viii.3.1 Charging . . . 218

viii.3.2 Discharging . . . 219

viii.4 System Analysis . . . 222

viii.4.1 System Performance . . . 222

viii.4.2 Cost Estimation . . . 224

viii.5 Conclusion . . . 225

viii.6 References . . . 225

(15)

©2021 by

Stefan Haller

LIST OF FIGURES

1.1 Simulated efficiency map of the next generation 48-phase low-voltage high-current (LVHC) motor with peak efficiency at the lower to medium power region. . . 13 1.2 Ragone plot comparison of various energy storage tech-

nologies with specific power versus energy density on cell level. . . 14 1.3 Prototype of a reconfigurable battery pack with parallel

to series connection switching for 48 V discharging and 400 V charging. . . 16 2.1 Overview of a principle low-voltage drivetrain. . . 25 3.1 Overview of included publications and progression to-

wards a 48 volt high-current machine. . . 27 4.1 Square-wave-like EMF of the first prototype machine at

300 rpm with 3 excited coils under no-load condition. . . 41 6.1 Sectional view of prototype II with mounted power elec-

tronics and VCC ring in the test-stand configuration with support bearing. . . 50 6.2 One of 13 identical preformed stator coils of prototype II. 51 6.3 13-phase MOSFET commutator, tested up to 600 A contin-

uous current. . . 52 6.4 Example of a 13-phase high-current PCB design with

aluminum core for a lab-winding machine. . . 53 6.5 Sectional view of the current design stage of the next

generation 48 V high-power machine with mounted power electronics. . . 55

(16)

©2021 by

Stefan Haller

LIST OF TABLES

2.1 Comparison of low-voltage and high-voltage drive system properties. . . 22 6.1 Design guideline comparison of the first prototype and

the second improved version, based on the results of the first prototype. . . 49

(17)

©2021 by

Stefan Haller

LIST OF PUBLICATIONS

This thesis is based on the following publications, attached in Part II.

Paper I

Air-Gap Flux Density Measurement System for Verification of Permanent Magnet Motor FEM Model

S. Haller, P. Cheng, B. Oelmann

IEEE Proceedings of Industrial Electronics 2015

Vol. 41, pp. 445–450 . . . 75 Paper II

Investigation of a 2 V 1.1 kW MOSFET Commutated DC Motor

IEEE Proceedings of Power Electronics and Motion Control 2016 Vol. 17, pp. 586–593 . . . 91 Paper III

Initial Characterization of a 2 V 1.1 kW MOSFET Commutated DC Motor

IEEE Proceedings of Industrial Electronics 2016

Vol. 42, pp. 4287–4292 . . . 113 Paper IV

A 2.5 V 600 A MOSFET-Based DC Traction Motor S. Haller, P. Cheng, B. Oelmann

IEEE Proceedings of Industrial Technology 2019

Vol. 20, pp. 213–218 . . . 129

(18)

©2021 by

Stefan Haller

Paper V

Rebalancing of Phase Current Shift Caused by Armature Reaction in Multiphase Single-Turn Concentrated Winding Machines

S. Haller, P. Cheng, K. Bertilsson

Submitted to IEEE Access . . . 147 Paper VI

CPLD and dsPIC Hybrid-Controller for Converter Prototyping driving a Reconfigurable Transformer Phase-Shifted Full- Bridge

S. Haller, M. Abu Bakar, K. Bertilsson

Proceedings of Int. Conference PCIM Europe digital days 2020 pp. 1552–1558 . . . 173 Paper VII

Multi-Phase Winding with In-Conductor Direct Cooling Ca- pability for a 48 V Traction Drive Design

S. Haller, J. Persson, P. Cheng, K. Bertilsson

IEEE Proceedings of Int. Conference on Electrical Machines 2020 Vol. 1, pp. 2118–2124 . . . 189 Paper VIII

Reconfigurable Battery for Charging 48 V EVs in High-Voltage Infrastructure

S. Haller, F. Alam, K. Bertilsson

In manuscript . . . 209

(19)

©2021 by

Stefan Haller

List of Publications

Publications not Included

The following publications by the author are not included in this thesis.

Contribution of leakage flux to the total losses in transformers with magnetic shunt

M. Abu Bakar, S. Haller, K. Bertilsson International Journal of Electronics, 2020 DOI 10.1080/00207217.2020.1793404

State of Art of Designing Power Electronics Converter for Low Voltage Motor Drives for Electric Vehicle

M. Das, S. Barg, S. Haller, M. Abu Bakar, A. Rezaee, K. Bertilsson IEEE International Conference on Power Electronics, Smart Grid and Re- newable Energy (PESGRE 2020), pp. 1–6

DOI 10.1109/PESGRE45664.2020.9070438

Assembling surface mounted components on ink-jet printed double sided paper circuit board

H. Andersson, A. Manuilskiy, S. Haller, M. Hummelgård, J. Sidén, C. Hummelgård, H. Olin and H. Nilsson

Nanotechnology, Vol. 25, ss. Art. no. 094002, 2014 DOI 10.1088/0957-4484/25/9/094002

A ZVS Half Bridge DC-DC Converter in MHz Frequency Region using Novel Hybrid Power Transformer

H.B. Kotte, R. Ambatipudi, S. Haller and K. Bertilsson

International Conference for Power Electronics, Intelligent Motion, Renew- able Energy and Energy Management (PCIM 2012), pp. 399–406

ISBN 978-3-8007-3431-3

System of nano-silver inkjet printed memory cards and PC card reader and programmer

H. Andersson, A. Rusu, A. Manuilskiy, S. Haller, S. Ayöz and H. Nilsson Microelectronics Journal, Vol. 42: 1, ss. 21–27, 2011

DOI 10.1016/j.mejo.2010.09.008

(20)

©2021 by

Stefan Haller

A study of IGBT rupture phenomenon in medium frequency resistance welding machine

J. Saleem, A. Majid, S. Haller and K. Bertilsson

International Aegean Conference on Electrical Machines and Power Elec- tronics and Electromotion, Joint Conference (ACEMP 2011), pp. 236–239 DOI 10.1109/ACEMP.2011.6490602

Analysis of feedback in converter using coreless printed circuit board transformer

A. Majid, J. Saleem, H.B. Kotte, R. Ambatipudi, S. Haller and K. Bertils- son

International Aegean Conference on Electrical Machines and Power Elec- tronics and Electromotion, Joint Conference (ACEMP 2011), pp. 601–604 DOI 10.1109/ACEMP.2011.6490667

High Frequency Half-Bridge Converter using Multilayered Core- less Printed Circuit Board Step-Down Power Transformer

A. Majid, H.B. Kotte, J. Saleem, R. Ambatipudi, S. Haller, and K. Bertils- son

8th International Conference on Power Electronics - ECCE Asia (ICPE 2011), pp. 1177–1181

DOI 10.1109/ICPE.2011.5944712

(21)

©2021 by

Stefan Haller

Part I

Introduction

(22)

©2021 by

Stefan

Haller

(23)

©2021 by

Stefan Haller

Chapter 1

OVERVIEW

The transition to e-mobility is a big challenge in terms of energy storage, fast energy delivery at charging stations, cost and drivetrain efficiency. Rarely discussed, but of major importance, is system safety.

In automotive engineering, voltages up to 60 V DC and 30 V AC are classified as low-voltage (LV) and handled according to the safety extra-low voltage (SELV) regulations. Voltages above these limits are classified as high-voltage (HV), requiring additional precautions [1] to avoid injury or even lethal accidents. To fulfill the regulations, double isolation of the high-voltage components and cables, for example, is mandatory. The orange double-isolated high-voltage cables may even require shielding. Additionally, isolation monitoring and 2-pole safety circuit breakers are needed to disconnect the high-voltage source [2]. The HV source is usually a battery. The nominal voltage level of current battery electric vehicles (BEVs) is higher than 320 V [3,4] and therefore classified as high-voltage. In addition to the common 400 V class, manufacturers are moving towards even higher voltages. The world’s first production car with an 800 V system was the Koenigsegg Regera, a high-performance sports car, which became available in 2017 [5]. In 2019, the first regular sports car, the Porsche Taycan, was launched [6]. It is the first mass production electric vehicle (EV) in the 800 V class.

In contrast, SELV rated systems do not require double isolation, and the interruption of a single pole is sufficient to disconnect the battery. SELV systems even allow the use of structural parts for energy transfer, like the well-established utilization of the vehicle’s chassis as ground.

The energy storage devices for e-mobility applications are currently the most sensitive and costly part of the drivetrain. Efficient use of this resource by drivetrains is key. By increasing system efficiency, driving range can be extended for a given battery capacity [4].

To find novel solutions for these challenges, pioneering ideas that break with principles applied for more than a century have to be

(24)

©2021 by

Stefan Haller

openly discussed. Recent developments in modern power electronics allow totally different drive concepts than those widely used only a decade ago. This thesis describes the journey towards a low-voltage, high-current DC drive system. It is an inherently safe, efficient and redundant pioneering concept for 48 V battery electric vehicles.

1.1 Traction Drives for EVs

Traction drives for EVs have to fulfill certain requirements. They need to be robust to survive in harsh environments where they are exposed to vibrations and both fast and slow temperature changes. Traction drives should deliver peak torque for acceleration and deceleration and thus need to handle fast current transients throughout their entire life span. They should furthermore be efficient, maintenance free and safe in case of failure or damage. The following sections discuss some novel ideas on how a LVHC drive can achieve this.

1.1.1 Resistive Losses

More than 50 % of the losses in today’s most frequently used electric motors, the squirrel cage induction (SCI) and permanent magnet (PM) motor, are resistive losses [7], [8]. The exact distribution of the losses depends on the operating conditions and motor design. A considerable amount of research on well-known drive architectures has been dedicated to minimizing the losses for a given application, such as driving cycles of EVs [9–11]. The resistive losses nonetheless dominate beyond a certain load of the electrical machine [12].

The winding layout and configuration is of major importance in minimizing resistive losses, also known as copper losses or 𝐼²𝑅 losses.

End-windings, also called end-turns or winding head, do not produce any considerable amount of net-torque and thus mainly contribute to the copper losses. The larger the usable cross-sectional area of the conductor and the more compact the end-winding layout, the lower the resistive losses. This also applies to the stator winding part inside the slots. Here, the winding layout and copper fill-factor is of importance. The copper fill factor is the ratio between the slot opening area of the machine’s electrical steel lamination and the area

(25)

©2021 by

Stefan Haller

1.1.2. Machine Cooling

of copper or aluminum conductor, which is fitted inside the slot. The general aim is to increase the fill factor as much as is practically and economically possible. The conductor isolation, fixating assembly and resin to increase the thermal conductivity are considered passive material. The electrical steel and the conductor inside the slot to build the machine’s winding are active material, as they actively contribute to the machine’s torque generation. In terms of PM machines, the magnets are also counted as active material.

The random-winding, by far the most common winding configuration, allows a copper fill factor of up to 0.4–0.5. Special winding techniques, such as needle windings, allow copper fill factors of up to 0.65 [13]. More complex winding techniques, such as concentrated pre-formed windings, further increase the copper fill factor. Fill factors of up to 0.8 have been reported for larger machines, such as a 25 kW 50 V PM machine [14]. In general, the larger the electrical machine, the easier it is to achieve a high fill factor. This becomes more of a challenge for compact high-power traction machines with small slots. For these machines, the hairpin winding technology is more frequently used [15]. Pre-formed rectangular hairpin-shaped wires are inserted in the slots and welded together. These windings provide a high fill factor and low thermal resistance to the stator core. The prototype machine presented in Papers IV and V use similarly shaped coils.

1.1.2 Machine Cooling

Both the first and second LVHC prototype machines are air cooled, like most industrial motors. Traction machines instead require a higher power density to reduce weight and size. A major portion of the machine’s losses are generated inside the machine’s stator winding.

Heat extraction close to or inside the winding is therefore a key factor in achieving high-power density. To cool these machines, various forced liquid or air-cooling techniques can be used.

Stator housings with a water jacket are easy to integrate but extract the heat through the machine’s stator lamination. Cooling channels embedded in the stator lamination can increase heat extraction capability, but still depend on low thermal resistance of the winding to

(26)

©2021 by

Stefan Haller

the stator. Potting the winding in the stator slots increases mechanical stability and reduces thermal resistance. With indirect slot cooling, the cooling channels are embedded in the slot, together with the winding [16–18]. This further reduces the thermal path between the winding and the cooling medium.

The larger the cross-sectional area of the coil conductors, the easier it is to use hollow or pre-formed conductors, which allow direct inner cooling. Direct conductor cooling is prevalent in very high power electrical machines, for example electrical generators at power plants.

Interest is growing in this form of cooling, to increase the power density of traction machines. For high-voltage machines, oil-cooled conducts are used [19]. With laminated windings, forced in-stator air cooling was demonstrated by Reinap et al. [20]. Compressed air or oil cooling with hollow or profiled conductors have been investigated by Reinap et al. [21]. They have reported current densities of up to 50 A/mm²with oil cooling at 40° inlet temperature [22].

For a multi-phase, low-voltage W-shaped winding with direct in- conductor cooling, using ethylene glycol water 50/50 (EGW50/50) at 65° inlet temperature,a current density of 49.5 A/mm²was reached, as shown in Paper VII. The use of a parallel hydraulic connection provides increased cooling capability in combination with low pressure. In- conductor cooling also eliminates the need for additional end-winding cooling.

Gai et al. [23] provide a comprehensive overview of common cooling solutions for automotive traction drives. They include indirect rotor cooling through the shaft, as used in Tesla’s induction machine in the early model S, and wet cooling with a fully flooded machine, as in Tesla’s model 3. Spray cooling of rotor and end-winding are also discussed.

1.1.3 Machine Faults

Considerable research has been conducted in different areas of electrical machines to determine fault mechanisms and develop fault prediction methods. According to studies on industrial motors [24], the most common induction motor (IM) faults were bearing faults followed by stator faults [25,26]. Almost 37% of all faults on industrial

(27)

©2021 by

Stefan Haller

1.2. MOSFETs or IGBTs for Traction Drives

machines were stator faults. More than 90 % of these stator faults were related to stator winding damage, of which the most prevalent was the stator inter-turn fault (SITF). The authors in [25] point out that traction motors are prone to SITF more than the industrial motors mentioned above. This is due to their high power density, very compact construction and extreme operating conditions. These drives are expected to work in harsh environments, are prone to fast as well as slow temperature changes, external as well as internal vibrations and fast current transients. Personal safety is furthermore the most important factor in vehicles and therefore a measure for traction drives.

Electrical failures, such as SITF, cause large circulating currents inside the damaged winding, which could cause very rapid deceleration of the vehicle. These circulating currents lead to local hot spots, which further degrade the isolation. Ultimately, faults that started as SITF result in complete failure of the winding as turn-to-ground and phase-to-phase faults as well as open-circuit faults [26,27].

PM machines are favored for traction drives due to their higher efficiency compared to SCI motors. Electrical failures with these type of machines are, however, more severe compared to SCI motor failures.

This is caused by the same effect, which makes them more efficient:

the nature of magnets to maintain a permanent magnetic flux. Once the rotor is spinning, an electromotive force (EMF) is generated that can hardly be counteracted. In case of a SITF, a spinning PM rotor will generate a high level of circulating currents in the shorted winding.

This generates a reverse magnetic field, which opposes the rotor’s field. In combination with high magnet temperatures, this could lead to permanent demagnetization and destruction of the rotor [28,29].

1.2 MOSFETs or IGBTs for Traction Drives

Until now, commercial high-voltage traction drive inverters commonly use insulated-gate bipolar transistors (IGBTs), as in the Nissan Leaf, Toyota Prius or Tesla Model S [30–32]. The development of the MOS- FET technology in recent years will lead to a progressive replacement of the IGBTs in traction drive applications.

(28)

©2021 by

Stefan Haller

MOSFET-based inverters have the advantage of significant loss reduction, and can thus increase the overall energy efficiency of the drivetrain [32]. The impact on the inverter losses by replacing commercial IGBTs-based traction drive inverters with silicon carbide (SIC) MOSFET-based ones has been discussed by Kim et al. [33].

A loss reduction on the inverter by approximately a factor of two was reached. The technology replacement has already started: to the author’s knowledge, the Tesla Model 3 is the first mass-produced BEV using SIC-MOSFETs in the traction drive inverter. More specifically, 24 SIC-MOSFETs from ST Microelectronics with 650 V rating are used [34].

While SIC-MOSFETs are designed for high voltages, commonly 1200 V, progress is also being made in the development of low-voltage silicon (SI)-MOSFETs. For example, the used 25 V IRF6718L2 MOSFET has a typical 𝑅^𝐷𝑆(𝑂𝑁)of 0.5 mΩ and can handle up to 270 A continuous drain current. The recent IRL7472L1 can handle up to 40 V with decreased typical 𝑅^𝐷𝑆(𝑂𝑁) of 0.34 mΩ and up to 375 A continuous drain current in the same package of just 9.1 mm × 7 mm × 0.7 mm.

The increased efficiency of the MOSFET technology compared to IGBTs, along with lower device costs and reduced cooling requirements, enables new machine designs. The aforementioned low-voltage MOSFETs are off-the-shelf components and have a good price performance ratio due to economy of scale.

As a result, efficient low-voltage drives that use a large number of switches are feasible, as in the presented LVHC drive or related Intelligent Stator Cage Drive (ISCAD) drive [35,36]. The advantages compared to high-voltage drives are the inherent electrical safety and reliability due to the use of many parallel current paths. A low-voltage drivetrain provides increased system efficiency at reduced complexity of the battery pack. This reduces cost and increases driving range.

1.3 Energy Storage and Delivery

While there is good understanding of HV motors and drives, energy storage technology is still the limiting factor of EVs. Batteries are by far the most common energy storage device in these vehicles. While the knowledge of and experience with lead-acid battery usage in

(29)

©2021 by

Stefan Haller

1.3.1. Lithium-Ion Technology

vehicles is considerable, their energy density is significantly lower compared to lithium-ion (Li-Ion)-based batteries [37,38]. As space and permitted weight are limited, battery technologies with a higher energy density need to be used in cars, such as nickel–metal hydride (NiMH) and today’s Li-Ion-based batteries. The use of the Li-Ion technology in EVs is fairly new though, and many challenges remain that require further research. Some of these are briefly discussed in the following sections.

1.3.1 Lithium-Ion Technology

It is worth mentioning a few properties among others that are chal- lenging to the current lithium-ion battery (LIB) technology when used in EV applications. The term Li-Ion battery encompasses a group of different types of Li-Ion-based battery cells and packaging. The exact material combination and layout determines the battery properties in terms of power, energy, high- and low-temperature performance, safety and lifetime. The mechanical packaging of the cells also needs to be considered, in addition to the selected active material.

There is considerable experience in production of cylindrical cells, such as the 18650, which is 18 mm in diameter and 65 mm long. This cell type has been used for more than a decade in laptop batteries.

The recent 21700 cylindrical cell type provides about one third more capacity and is just slightly larger. It is used in the Tesla Model 3, for example. The cylindrical housing is solid, can withstand some internal pressure and provides mechanical protection of the cells. The lower packaging density of cylindrical cells compared to semi-soft pouch cells can be considered a disadvantage. While pouch cells provide a slightly higher energy density due to the usage of a case foil instead of a cylindrical can, they are mechanically more fragile. A third type, commonly used in EVs, is the prismatic cell. Its assembly process is similar to that of the cylindrical cell, but has a higher packaging density.

Li-Ion batteries cells swell slightly while they are cycled [39]. This mechanical displacement needs to be considered when choosing the type of battery cell, package and assembly of the pack. Prismatic and cylindrical cells with a dense containment have the advantage of

(30)

©2021 by

Stefan Haller

providing clearance for swelling [38].

In contrast to lead-acid batteries, the performance of Li-Ion-based batteries drastically decreases at extremely cold temperatures [40].

The usable energy density of lithium iron phosphate (LFP) batteries, for example, is reduced by more than 40 % at−20 °C [41]. In addition to the capacity reduction, the cell resistance will also rise due to the slower Li-Ion diffusion rate. While this reduces the performance of BEVs with a cold battery, it limits the use of LIBs to cold-crank internal combustion engines (ICEs). Thus, mild-hybrid vehicles often carry a 12 V lead-acid battery which delivers sufficient cold cranking amps (CCA).

While discharging at a limited C-rate is still possible, charging capability is limited or may even need to be prevented to avoid plating, which constantly degrades the usable cell capacity [42,43]. Different pre-heating techniques can be used to heat the cells to a more optimal operating temperature. A promising technique involves internal cell heating by applying alternating currents [44,45]. Within a couple of minutes, the cells can be uniformly pre-heated to temperatures above 0°.

1.3.2 Risk of Li-Ion-based Energy Storage Systems

Depending on the chemistry used, every series-connected LIB cell adds approximately 3.3 V to 3.7 V to the battery pack’s voltage. Parallel- connected cells increase the pack’s capacity and current handling capability. All parallel connected cells sharing the same voltage level inside the pack are further denoted as cell stack.

The high energy density of EV battery cells, with more than 680 Wh/L or 260 Wh/kg [37], require that safety precautions are taken to protect the battery from mechanical and electrical damage.

Battery safety precautions can be passive on the cell and pack levels, as well as active via the battery management system (BMS). The composition of the cell chemistry directly influences the cell’s passive safety. Increased passive cell safety lowers the voltage and thus energy density.

In addition, the material combination of today’s Li-Ion-based cells does not provide an integrated overcharge protection mechanism [38].

(31)

©2021 by

Stefan Haller

1.3.3. Cell Configurations

Consequently, an external BMS that monitors the temperature, voltage and state of charge (SOC) of every cell stack is required. The properties of BMS systems range from simple to more advanced features, from just passively protecting the cells from overcharge or deep discharge to active cell level balancing, SOC monitoring, control of active cooling or heating, among others. EVs usually use an advanced BMS to ensure safety as well as active thermal and endurance management for the costly battery pack.

1.3.3 Cell Configurations

The current 400 V class of BEVs often use 96s𝑁p cell configurations [46], in other words 96 series connected battery cells, leading to a pack voltage of approximately 350 V. The pack’s capacity is either adjusted by the amount 𝑁 of parallel connected cells, or by slightly altering the number of series connected cells. For example, a Tesla Model S with a 85 kWh battery uses a 96s74p configuration [46] with 96 series and 74 parallel connected cells. For Volkswagen’s Modularer E-Antriebs-Baukasten (MEB), it is assumed that 96s2p, 108s2p and 96s3p configurations with 24 cells in each module are initially used [47]. The resulting nominal voltage is between 350 V to 410 V, which is classified as high-voltage.

Every series connected cell stack requires a dedicated channel on the BMS. The higher the battery pack’s voltage, the more channels on the BMS are required. Furthermore, due to voltages above the SELV limit of 60 V, the BMS requires more complex voltage and current sensing techniques.

Battery cells deviate slightly from their nominal capacity as result of the production spread. This spread increases during cell aging.

The string of series connected battery cells behaves like a chain. It is as capable as their weakest link. Each cell faces the same current and the cell with the lowest capacity determines the string’s capacity.

To strengthen the link, additional cells can be added in parallel to each link. The more parallel connections are used, the greater the robustness of the string against single cell faults and production spreads.

By reconfiguring a 400 V class battery pack from 96s74p to 48 V

(32)

©2021 by

Stefan Haller

with 13s547p, Baumgard et al. [48] have shown that the pack’s capacity and lifetime could be slightly increased by approximately 1 % and 1.5 %, respectively. It should be noted that the number of cells and individual cell current within the battery pack does not alter. The benefits of such a 48 V battery pack are increased redundancy and intrinsic safety due to SELV level. Furthermore, the cost of the battery pack is reduced by limited isolation effort, simpler cell connection schema and a simplified BMS, which requires 86 % fewer channels.

1.4 Battery Cost and Drivetrain Efficiency

In 2012, the cell price ranged from 200e/kWh to 500 e/kWh [38] and was estimated as 130e/kWh to 170 e/kWh in 2017 [49]. By the end of 2019, the battery pack’s average price had fallen below 160 $/kWh [50].

Even though cell and pack prices are constantly decreasing, the battery remains the most costly part of a BEV drivetrain. This fact emphasizes the importance of efficient use of the installed battery capacity to increase the BEV’s range.

The drivetrain’s peak power is only occasionally required on the road. In fact, under normal driving conditions below top speed, it is only used during heavy acceleration and recuperation phases. To maintain the vehicle’s speed on a flat road, only a fraction of the peak power is required [4,12]. A drivetrain with a peak efficiency at lower to medium power regions therefore maximizes the BEV’s range. A declining efficiency at high to peak power operation is acceptable due to its minor impact on the range. Figure 1.1 shows a simulated torque-speed efficiency map of the current state of development of the next generation 48 V traction motor. The peak efficiency is reached at the lower to medium power range.

A comparison of a 360 V high-voltage drivetrain and a 48 V low- voltage drivetrain was presented by Patzak et al. [4,36] in a simulation- based study conducted with the new european driving cycle (NEDC) and worldwide harmonized light vehicles test cycle (WLTC). It was shown that an equivalent low-voltage system outperforms a high- voltage system at every operational point in terms of efficiency. The

(33)

©2021 by

Stefan Haller

1.5. Energy vs. Power Density

Figure 1.1: Simulated efficiency map of the next generation 48-phase LVHC motor with peak efficiency at the lower to medium power region.

efficiency in low to medium power demand regions improved by more than 10 %.

1.5 Energy vs. Power Density – Battery and Supercapacitor Today’s Li-Ion batteries provide among the highest energy density of batteries used for automotive applications. While a high energy density is crucial for the BEV’s range, it limits the batteries’ power density, charge and discharge rate and performance of the vehicle.

Li-Ion batteries are offered as two major types, high-energy and high- power, as shown in the ragone plot in Figure 1.2. By changing the cells’ internal layout and battery chemistry, their properties can be shifted between high power and high energy density. An overview of automotive battery technology was published by Budde-Meiwes et al.

[38], while Schmuch et al. [37] have comprehensively summarized performance, energy density and material costs of lithium-based automotive batteries of current EVs.

In addition to the good range ensured by high energy density batteries, BEVs require high power to deliver peak torque and allow efficient energy recuperation during braking. The regular acceleration and recuperation phases lead to constant cycling of energy storage.

Supercapacitors can be used to reduce cycling and peak power stress

(34)

©2021 by

Stefan Haller

20 40 60 80 100 120 140 160 180

00 10 100 1000 10000

200 Specific Energy at Cell Level (Wh/kg)

Specific Power at Cell Level (W/kg)

220 240 260

Pb Supercapacitors

NiCd NiMH

Li-Ion High Power

Li-Ion Very High Energy Pb (spiral wound)

NaNiCl₂ (Zebra)

Figure 1.2: Ragone plot comparison of various energy storage technologies with specific power versus energy density on cell level. Modified figure from [51].

on the battery. These provide among the highest power density and pulse efficiency, although their energy density is fairly low as shown in Figure 1.2. The combination of a high energy density battery as long-term storage and a high power density supercapacitor as short-term storage results in a hybrid energy storage system. Since the energy stored in a supercapacitor is quadratic to the voltage, a DC/DC converter between battery and capacitor is required to utilize it in an efficient way. This additional DC/DC converter allows precise control of the capacitor current and SOC. The concept of a hybridized battery was presented by Carter et al. [52]. A resonant DC/DC converter design for this application was presented by Arazi et al. [53]. It was found to be an efficient method to reduce peak battery currents and thus extend battery life.

1.5.1 Environmentally Friendly Supercapacitor

Commercial supercapacitors are generally expensive, flammable and possibly toxic. This is due to the use of organic electrolytes in their composition, which allows an operation voltage of approximately 2.7 V. As an alternative, low-cost environmentally friendly nanographite coated paper-based supercapacitors can be used [54]. Their use of aqueous electrolytes limits the operating voltage to approximately 1 V. To reduce the balancing effort of the stacked low-voltage supercapacitors, it is favorable to use a low-voltage drivetrain.

(35)

©2021 by

Stefan Haller

1.6. Low-Voltage Charging

1.6 Low-Voltage Charging

Many different charging connectors and power ratings are currently used for EV charging. They can be split into two main groups: AC and DC charging systems. Some connectors and charging systems support both of them, like the IEC Type 2 or combined charging system (CCS) connector.

AC systems require an on-board charger that can be tailored to the battery pack’s capabilities. DC systems that are mainly used for fast charging have the converter integrated into the charging station.

The EV’s battery is directly charged via DC from the station, under supervision of the BMS.

To charge a 48 V drivetrain battery pack with 11 kW, a charge current of approximately 220 A is required. Since AC systems use an on-board charger, it can be closely mounted to the 48 V battery pack.

Greifelt et al. [55] presented such a 11 kW bidirectional SIC MOSFET charger. Additional effort is required to handle the high currents associated with DC fast-charging for low-voltage EVs. Investigations of the usage of high-temperature superconductors instead of copper or aluminum cabling at the charging station have been conducted by Greifelt et al. [56]. Their solution would theoretically support low-voltage charging up to 60 kW.

Instead of using superconductors, the author suggests using the existing high-voltage DC charging system. This could be achieved by on-board DC/DC point of load (POL) converters, integrated directly into the battery pack. An alternative concept with less power conver- sion is a dynamic reconfigurable battery pack, investigated in Paper VIII.

Battery packs for EVs are commonly constructed with multiple modules made of individual cells. During charging these modules can be connected in series to match the charger’s high-voltage. Once the charger is disconnected, the modules are switched to parallel connection to provide low-voltage to the drive. The measurement setup is currently being prepared to test the concept for series charging and parallel discharging with eight 48 V batteries. The batteries with the first printed circuit board (PCB) generation mounted on top for series to parallel switching are shown in Figure 1.3.

(36)

©2021 by

Stefan Haller

Figure 1.3: Prototype of a reconfigurable battery pack with parallel to series connection switching for 48 V discharging and 400 V charging.

Another fast charging solution may be derived from the concept of a contactless magnetic plug presented by Beddingfield et al. [57].

The authors suggest splitting the transformer into two parts, with the primary winding using 3.5 kV inside a magnetic plug and a 400 V secondary winding mounted inside the vehicle. The plug’s size should be comparable to a CHAdeMO 1.0 plug. With the secondary winding mounted inside the EV, it may be tailored to match a 48 V battery.

1.7 Related Low-Voltage Drives

The ISCAD from Molabo [36,48,58,59] is the most similar drive concept to the traction drive presented in this thesis. Both concepts propose a high copper or aluminum fill-factor, a redundant multi- phase electrical machine and MOSFET-based power electronics. The high fill-factor of the presented machines is achieved by single-turn windings, and for the next generation 48 V machine a winding with W- shaped coils. The ISCAD instead uses a stator cage, which can be seen as a half-turn winding design with common star connection. Both

(37)

©2021 by

Stefan Haller

1.7. Related Low-Voltage Drives

drives use on-stator mounted power electronics, which individually powers each stator phase.

The prototype machines presented in this thesis are PM machines with a square wave-like air-gap magnetomotive force (MMF). The commutation is done block-wise in today’s power electronics. The machines allow up to six simultaneously powered phases with unidirectional coil current. The armature reaction induces a phase current shift and jitter in the rotor position detection, as described in Paper V.

The machines were therefore always operated with five of 13 phases at a time. The unidirectional coil current simplified the power electronics design and enabled a terminal resistance of 0.23 mΩ. When powering only a fraction of the coils, however, the utilization of the machine’s active material is limited.

The ISCAD concept instead uses an SCI rotor and is driven in such a way that a sinusoidal air-gap MMF is generated. The stator current is bidirectional due to the use of a half-bridge configuration per phase.

A 48 V 110 kW 42-phase ISCAD machine in a fully functional car is shown by Runde et al. [59]. The first series produced ISCAD V50 [58] is a permanent magnet assisted synchronous reluctance (PMaSR) machine delivering up to 80 kW.

The concept of a 48 V concentrated single-turn bar windings machine reaching a high fill-factor is already described by Endert et al.

[14,60]. The machine uses a classic 3-phase stator winding instead of one independent coil per phase, however, and thus lacks redundancy in its design in case of a single phase failure.

(38)

©2021 by

Stefan

Haller

(39)

©2021 by

Stefan Haller

Chapter 2

PROBLEM FORMULATION AND OBJECTIVE

As described in the overview in Chapter 1 a fully low-voltage drivetrain seems beneficial in many aspects. It is intrinsically safe, no high voltage is created at any point and SELV limits are always maintained. Current BEV traction drives favor the 400 V to 800 V class. By regulation, this requires specially trained service personnel, and in case of a car accident, additional safety precautions are needed. Low-voltage drivetrains do not require these efforts.

Low-voltage drives have potentially greater efficiency due to the high copper fill-factor and the use of MOSFET-based power electronics.

Redundancy against single phase faults is achieved via the massive multi-phase design of the electrical machine and power electronics.

The most common electrical machine fault, the SITF, can be eliminated by a single-turn or half-turn winding. The stator side cooling of such windings is greatly enhanced by the high copper fill-factor and thinner coil isolation. Due to the larger cross-sectional area of the coil wires or bars, hollow conductors for direct inner cooling can be used. This drastically increases the continuous power density of the machine.

The low-voltage coils are connected electrically and hydraulically in parallel, which reduces the required coil coolant flow rate and absolute pressure.

The battery pack’s complexity and cost is reduced by using a low- voltage massive parallel cell configuration instead of a high-voltage massive series configuration. A reconfigurable battery pack for low- voltage discharging and high-voltage fast charging is currently being investigated. To reduce short-term storage cycles and current spikes to the battery pack, supercapacitors can be added that are low-voltage in nature.

When this work was initiated prior to the Diesel scandal in 2015, there was no major interest in low-voltage electrical machines for battery-powered traction applications. Research activity on electrical

(40)

©2021 by

Stefan Haller

transportation has drastically increased since then, however.

2.1 Towards Low-Voltage, High-Current

To generate a rotating air-gap MMF of desired strength, a certain number of stator coil ampere-turns is required. Ampere-turns are a linear function of the coil current and number of turns. More simply explained, a coil with 1 A and 100 turns as well as one with 100 A and 1 turn generate an air-gap MMF of similar strength.

For more than a century, electrical machines were designed to operate with higher voltage and less current. This mindset is derived from Ohm’s law, since the power 𝑃 scales with the square of the current 𝐼,

𝑃 = 𝐼²·𝑅 (2.1)

with 𝑅 as opposing resistance. Nevertheless, it can be beneficial to use low-voltage, depending on the power source, sink and application.

To maintain the same power, the current needs to be consequently increased.

To use an analogy: In many parts of the world, residential build- ings are supplied with up to 400 V, while most home appliances with external power supplies stay below 60 V. Most battery-powered devices use voltages below 60 V as well, and for good reasons: safety, costs, efficiency and usability. Most people would agree that it hardly makes sense to construct a 230 V pocket light to be able to use off- the-shelf household light sources just to reduce the conduction losses due to decreased current. While this is an extreme example, the same principle should apply to BEVs and hybrid electric vehicles (HEVs) as well.

In the author’s opinion, it does not make sense to construct a high-voltage electric vehicle utilizing the same low-voltage battery technology as that used in mobile phones. The motive for using high-voltage might be derived from the existence and knowledge of conventional industry-approved 400 V class drive systems and time to market-related pressures. Yet without considering the complete system, an optimal solution can hardly be found.

First attempts to raise the voltage level above 12 V for passenger vehicles were discussed in the late 1990s. This 42 V system actually

(41)

©2021 by

Stefan Haller

2.2. Problem Formulation

used 36 V batteries and was introduced to replace the existing 12 V board grid [61]. For various reasons it never gained market share and was only used in two car models. A few years ago, a new attempt at a 48 V system was made, this time as an addition to the existing 12 V board grid. The voltage range for the 48 V system was standardized through the VDA320 [61] in 2014. The stationary voltage is always kept within the SELV limits, below 60 V DC.

To be able to meet the upcoming requirements on CO2emission reduction, more and more manufacturers are equipping their ICE- based cars with an additional 48 V mild hybrid system. In addition to emission reduction, additional comfort features like boosting and e-creeping are possible.

P0 belt starter generator (BSG) machines with 10 kW and 48 V are commonly available. Original Equipment Manufacturers (OEMs) are already offering integration support of their 48 V drives with power levels of 20 kW to 30 kW [62], ranging from belt-connected P0 to electric all-wheel drive P4 solutions. Some OEMs have stated that up to 50 kW will be available in future generations of automotive 48 V-based machines. Recently, Molabo released a 48 V drive with 80 kW peak and 50 kW continuous power for boats [58]. They also demonstrated a 48 V 110 kW 42-phase ISCAD machine in a fully functional car [59].

All this raises the question: why should a third, high-voltage system be used for EVs if the task can already be accomplished by a 48 V system?

2.2 Problem Formulation

A low-voltage drivetrain has many advantages but also several challenges compared to a high-voltage system. Table 2.1 compares some key properties of the proposed drivetrain against a HV one.

Over the last century, a solid understanding of, and extensive experience with high-voltage machines has been obtained. For multiphase LVHC machines, however, experience and research is very limited.

This thesis provides an introduction into the field of low-voltage high- current machines, their advantages and limitations. The following research questions were raised and have been answered by this work.

(42)

©2021 by

Stefan Haller

Table 2.1: Comparison of low-voltage and high-voltage drive system properties.

Property HV System LV System

battery cell setup few parallel many parallel BMS channels many isolated ≤ 13 unisolated drive origin industrial HV LV battery

redundancy limited/none redundant, modular

electrical hazard lethal safe

drivetrain efficiency good better challenges

safety low resistance

isolation LV fast charging redundancy key-component de-

velopment

2.2.1 LVHC Machines with Novel MOSFETs

When considering the recent advances made in high-current MOSFETs, is it possible to use these switches to construct a compact multi-phase LVHC machine with integrated power electronics? What would be the advantages and limitations of such a design?

These main research questions are addressed by the studies presented in Papers I–III. To answer them, a first prototype machine was constructed, and in doing so the following design-specific questions have been addressed:

• Which MOSFETs can be used to achieve minimal 𝑅^𝐷𝑆(𝑂𝑁)?

• What is the simplest machine topology for an initial proof-of- concept?

• What power electronics topology can be integrated with suitable effort, and what trade-offs need to be made?

• What controller should be used?

• How should machine and power electronics be assembled to provide a detachable, low resistance connection?

(43)

©2021 by

Stefan Haller

2.2.2. Improvement and Limits of LVHC Machines

• How should the FEM model be verified for further design improvements?

2.2.2 Improvement and Limits of LVHC Machines

The first prototype demonstrated that LVHC machines are feasible with today’s power electronics.

This leads to the question: what efficiency improvements can be made while keeping the previous machine topology and power electronics? What limits exist and what issues arise that require further investigation?

To answer these questions, a second prototype was constructed, and is discussed in Papers VI and V.

2.2.3 Components for 48 V LVHC Machines

Both prototype machines served to validate the concept of multi-phase LVHC machines. As one of their limitations, the lag of individual phase current control and unidirectional current flow was identified.

This led to the question of what type of multi-phase pulse-width modu- lation (PWM) controller can be used.Paper VI presents the study of a potential multi-phase PWM controller.

Another obstacle for a 48 V traction drive is the limited cooling capability of the presented prototype machines. The question of how the stator winding can be cooled efficiently is answered in the study published in Paper VII.

To investigate a 48 V fast-charging concept for traction batteries, a reconfigurable battery pack for high-voltage charging and low-voltage driving is studied in Paper VIII.

2.3 Objective

The main objective of this work is to determine if multi-phase LVHC drives with integrated power electronics can be built by using today’s high-current MOSFETs. This type of machine breaks with the con- vention of reducing the current by increasing the voltage to reduce conduction losses. These machines can theoretically be built and would offer many advantages if sufficiently low path resistance could

(44)

©2021 by

Stefan Haller

be achieved to compensate for the 𝐼²𝑅 losses that scale quadratic with the current. Many practical details have to be considered, however, and there are inevitable compromises when building such a machine.

This work focuses on the practical implementation of multi-phase LVHC machines with integrated power electronics. It elucidates the solutions and limitations of low-voltage high-current, which are intrinsically safe and redundant by design. The first proof-of-concept machine confirmed that today’s high-current MOSFETs are not the limiting factor. To investigate the weaknesses of the concept and to develop an understanding of this type of machine and high-current power electronics, a second prototype was constructed. With the promising results and experience gained through this small-scale 600 A machine, the key components for a 48 V high-current fully electric drivetrain are now being investigated.

2.3.1 Limitations

The block diagram of a principle low-voltage drivetrain is shown in Figure 2.1. This thesis focuses on the LVHC machine with a MOSFET-based drive (green boxes). The reconfigurable battery is currently being investigated and constitutes future work. The LV and HV charging, as well as the LV supercapacitor and adjacent DC/DC converter are not included in this thesis.

This work does not claim to provide an in-depth analysis of electrical machines. Rather, this thesis provides an overview of, and motives for further investigation of low-voltage, high-current, battery- powered drives.

(45)

©2021 by

Stefan Haller

2.3.1. Limitations

Included in this thesis

LVHC machine MOSFET

drive LV mode

LV Charger

HV mode HV fast

Charger

Reconfigurable Li-Ion Battery

DC/DC converter LV Super-

capacitors

Figure 2.1: Overview of a principle low-voltage drivetrain. This thesis focuses on the LVHC machine and MOSFET drive (green boxes). A reconfigurable Li-Ion battery pack (blue boxes) is currently under development and constitutes future work. The parts in the yellow boxes are not included in this thesis.

(46)

©2021 by

Stefan

Haller

(47)

©2021 by

Stefan Haller

Chapter 3

PUBLICATIONS INCLUDED IN THIS THESIS

The thesis includes eight papers, found in Part II. Six of them have been published at the time of the thesis’ publication and one is in manuscript. Five of the papers are IEEE peer-reviewed publications.

An overview of the publications is illustrated in Figure 3.1.

Papers I–III focus on the LVHC proof-of-concept by designing and constructing the first prototype machine. Papers IV and V discuss the second prototype machine, its torque-speed efficiency measurement and armature reaction to its concentrated single-turn windings. With the LVHC concept proven to work, the next drivetrain generation aims to implement a 48 V system. The investigation of the required components has been initiated with studies for Papers VI to VIII.

A short summary of all papers and their contribution to the thesis, including the the author’s contribution, is found in this chapter.

Paper 1 Air-gap flux density

Paper 2 Motor In- vestigation

Paper 3 Characte-

rization Prototype

Machine 1

Paper 4 2.5V 600A

motor

Paper 5 Armature Reaction Prototype Machine 2

Paper 6 Controller

Paper 7 Cooling

Paper 8 Battery Next Gen- eration 48 V

Figure 3.1: Overview of included publications and progression towards a 48 volt high-current machine. Paper VIII is currently in manuscript and constitutes future work.

Towards Low-Voltage, High-Current: A pioneering drive concept for battery electric vehicles

©2021 by

Stefan Haller