EP/I038543/1
Leigh Murray
University of Warwick
1
 Objectives
 Facts
& Figures
 Outputs
2

To develop new EV technologies.

Meet challenges + opportunities facing the EV market.

Integrate electrical motor + power electronics
◦ reduce cost/weight and increase power density
◦ improve reliability of electrical power systems
◦ maintain manufacturability for a mass market
3

EPSRC-funded project: £3.1 M + £745k for demo projects

4-year project: Oct 2011-Sep 2015. Extension to Mar 2016

10 partners

6 research themes

3 technology demonstrators (started Oct 2013)

Low TRL (1-3) to support EV technology development
4
5

Experts in power electronics, electrical machines, and mechanical
engineering
Prof Phil Mawby
(Warwick)
Prof Andrew Forsyth
(Manchester)
Prof Phil Mellor
(Bristol)
Prof Volker Pickert
(Newcastle)
Prof Keith Pullen
(City)
Prof Mark Johnson
(Nottingham)
Prof Patrick Luk
(Cranfield)
Prof Emil Levi
(Liverpool John Moores)
Prof David Stone
(Sheffield)
Prof Andrew Cruden
(Southampton)
6
•
•
•
•
•
•
Car manufacturing companies
Semiconductor manufacturers
Component manufacturers
Product designers
Energy supplier
Consultancies
7
Demo 1: Integrated
Non-Rare-Earth High
Performance Drive
Demo 2: Integrated
Power Conversion for
Reduced EMI
Demo 3: An Integrated Onboard Battery Charger using a
Nine-phase Machine, with V2G
Capability
8
Objective: To grow 3C-SiC on Si wafers for high current and high voltage power devices. Not 4H-SiC
S
because too expensive.
Specific aims:

Create lateral MOSFET devices with:
o medium/high blocking voltages
o high current  permits high
torque, high acceleration
Finger vs circular structure: Take up space /(non) uniform
distribution of current.
Passivation Oxide
Source
Gate

Target = 1200V and 10A = future
requirement for higher performing EV
power trains.
Achieved:
• Lateral MOSFET with a channel mobility
of around 90 cm2/Vs which is much
higher compared to 4H-SiC.
•
So current density is expected to be high.
Al
Gate
Ti/Ni
Source
Ti/Ni
Drain
Drain
Passivation
Oxide
Fabricated
lateral MOSFET
Commercial 4HSiC: 40cm2/V.s
9
D
Challenge: To achieve high voltage with current material quality because…
•
Interface strain from lattice mismatch. 3C-SiC lattice structure is smaller than Si.
•
Si and SiC expand at different rates (Si expands faster than SiC when heated).
3C-SiC
on Si
3C-SiC
anti-phase
boundary
3C-SiC
Stacking
faults
Si
•
Strain is partially released through film curvature (macro view)
and defect formation (micro view).
•
Stress-induced defects formed in 3C-SiC can cause leakage
current.
Si
10
Integrated Non-Rare-Earth High Performance Drive
Total Funding: £311,982
Objectives:
• To develop a high performance ferrite
motor with full functional integration with
its converter.
• Mechanical and thermal integration of the
motor and the controller.
Theme 2: Design Tools
(Newcastle, City, Manchester)
Theme 4: Motors (Cranfield
and Newcastle)
11

Experiments  data  correlations fed into design of software (enable
thermal design of the electrical machine).

Rotate rotor at various speeds and apply forced airflow from a fan
blower.

Forced flow results in a more uniform heat transfer (stator to air).
For lower rotational speeds, more airflow causes heat transfer to increase.
For higher rotational speeds, more airflow may cause heat transfer at the
rim to decrease slightly.
No forced flow
Radial outflow: 8 g/s
Radial outflow of 19 g/s
12


Power module failure: cracks in solder layer and bond
wire lift-off.
Newcastle have integrated a cooling medium to a power
module to cool power chips individually.
Results:
 Temperature fluctuation (∆T) reduced by 10 oC by using
liquid-metal material  has potential to double chip
lifetime compared to conventionally cooled power
modules.
Pipes containing the
liquid-metal coolant.
Power supply for the
internal pump (no
mechanical rotational
component).
160
140
o
Tjunction ( C)
120
Power module and cooling circuit
Red – chip temp. cycle new module
Blue – chip temp. cycle traditional module
100
80
60
Proposed with periodic flowrate
Conventional with constant flowrate
40
0
5
10
15
20
Time (s)
25
30
35
40
13

Rare-earth PM in-wheel machine (speed <1000 rpm)
◦ Relatively new concept of drive train.
◦ PE integrated inside the stator to directly turn the
wheels.
◦ No gearbox or transmission system, so space for battery.
◦ Mass of in-wheel motors can’t be large - stability.
Energy Source
In-wheel Motor
◦ Disadvantages of rare-earth: high-temperature
demagnetization; expensive; restricted supply.
Motor Controller

Ferrite-based PM machine (speed ~5000 rpm)
◦ Working temperature 100⁰C higher than rare-earth PM.
◦ Disadvantages: low-temp. demagnetization; fragility;
low-energy density (1/10 of rare earth).
14
Rotor
 Used multi-physics optimization to simulate
structure.
 Decided on a small rotor diameter and pole
number (8) to simplify the rotor structure and
safeguard rotor integrity against high centrifugal
force. Max speed: 20,000 rpm.
Lamination stress
reduced to <300MPa
at 20,000 rpm.
Stator

Compared 4 stator configurations with different
slot numbers.

Assessed affect on torque output, torque ripple,
and demagnetization.

Used 3-D electromagnetic FEA determined the
final optimal design.

Stator with 48 slots and 156mm diameter.
30 slots
42 slots
36 slots
48 slots
15
•
•
•
High-strength, support pins used to reduce stress in the lamination steel.
Optimal rotor has eight poles and is 95.4mm diameter.
Employ Nippon Steel 0.35mm lamination 35H250 with 420MPa tensile strength
to minimize rotor core loss.
Rotor right cover
Rotor left
cover
Pin holders
Rotor shaft tie pin
Lamination
with pin
holders
Lamination support pins
16
Employ Nippon Steel 0.2mm lamination 20HTH1200 to minimize the stator core loss at high rotational speed.
Stator lamination with
windings
Partially assembled stator
Aluminium casing
DC capacitor and power module
(converter and DSP inside) mounted
on Al casing on stator
17

Rare-earth motor (with Demo 1 specs)
Demo 1: Ferrite PM motor
Material weight
• Lamination: 9.0 kg
• Cu wire: 1.2 kg
• Rare earth PM: 0.75 kg
Total = ~11 kg
• Lamination: 12.0 kg
• Cu wire: 1.7 kg
• Ferrite PM: 1.0 kg
Total = ~15 kg
Cost (inc.
manufacture)
High. Rare-earth PM = £50/kg
Low.
Ferrite PM = £7/kg
Nominal rating: 20 kW at
10,000 rpm (max. 20,000
rpm).
Cross-sectional view of
the final assembly
Overview of the final assembly.
Integration of controller board on top.
18
Integrated Power Conversion for Reduced EMI
Total Funding: £310,661
Objectives:
• Save weight and volume through having a sealed
enclosure and shared cooling circuit.
• Reduce electromagnetic emissions by using filters
and Al enclosure.
• Reduce the passive component size to reduce
converter size and increase power density.
Theme 3: Packaging
(Nottingham)
Theme 6: Passives (Bristol,
Manchester, Sheffield)
19


Passive components represent ~50% of the mass of a power converter.
Objective: investigate technologies to deliver compact and lightweight
wound components, and to increase heat removal from them.
Heat removal:

Encase in a composite potting compound (epoxy).

Adding Al2O3 powder as filler material can improve thermal conductivity.
Epoxy pumped in
Toroidal inductor
With filler get 20% more heat removal
from inductor.
•
E-core inductor
Potted inductor with
thermal sensors
Heat extraction better with e-cores: end windings
potted, and can mount on cold plate.
20
High-fidelity modelling to find best combination of coil and core dimensions. Consider
•
electromagnetic performance
•
AC losses from windings and core losses
•
thermal behaviour simultaneously
VESI inductor requirements:
- 80 μH inductance
- 200 A rated current
- 400 Hz operating frequency
Winding power loss [W]
250
Al conductors 20 C
Al conductors 40 C
Al conductors 60 C
Al conductors 80 C
Cu conductors 20 C
Cu conductors 40 C
Cu conductors 60 C
Cu conductors 80 C
200
150
100
Al
Cu
50
0
Al windings in e-core. End
windings to be potted in epoxy.
1
Inductor
Design
Energy density for
theoretical mass of
2.5 kg
Commercial
0.1-0.2 J/kg
Publications
0.2-0.8 J/kg
VESI
1.2 J/kg
2
3
4
Frequency [kHz]
5
6
21


Alternative ways to package chips on a substrate instead of wire-bonding and soldering techniques.
Employ integrated, modular assemblies, where electromagnetic, thermal and mechanical functions
are treated together.
Multi-layer flex
connect with
sintered SiC JFETs
Sintering
process
AlN substrate with Ag nano-paste
4-channel
integrated
liquid
cooler
Demo 3
uses 3
power
modules
to create
a 9-phase
converter
Integrated
gate drive
circuit
Integrated CM and DM input filter
Final module (12.8 x 16.3 x 3.3 cm)
22
•
•
By mounting the main power conversion elements within a single enclosure, with appropriate power and
signal filtering, the associated EMI challenges can be eliminated.
Integration  at least a 20% reduction in volume and weight.
COLD PLATE LAYOUT
DC-DC
converter
FPGA control board
Power module for traction
motor
Bi-directional DC-DC
converter interface
to 30kW supercapacitor buffer
store
23
An Integrated On-board Battery Charger using a Nine-phase Machine,
with V2G Capability
Total Funding: £269,437
Objectives:
• To develop a working prototype system of an
onboard charger with bi-directional power flow .
• To develop a high-power-density power
converter, based on WBG devices.
• To design control software for charging and V2G
functionality.
Theme 3: Packaging
(Nottingham)
Theme 5: Converters (Manchester, LJMU,
Newcastle, Southampton)
24




Investigate integrated on-board charger with bidirectional power flow for battery charging and V2G
operation.
Can use single-phase (slow), 3-phase, and multiphase charging (fast).
No separate charger, instead RE-USE the pre-existing magnetic components and inverter installed for
driving mode.
No torque produced while charging.
Advantages:
 Fewer new elements  lower cost
 Lower weight  faster vehicle
 Less space needed  smaller vehicle
 Can use any type of power socket
 V2G operation  helps with providing stored electricity to the grid
25

Concept is applied to 5,6, and to 9-phase configurations.
iag
three-phase
grid
vag iag
+
vbg ibg
+
vcg icg
+
EMI
filter
(onboard )
S1
idc
five-phase
machine
b
ibg
c
S2
icg
a
hardware
reconfig.
iL
ic
icg /2
ic
ibg /2
ib
iag
ia
ibg /2
ie
icg /2
id
vc
vb
va
vdc
C
BAT
ve
vd
d
e
• Switches open  grid is connected and can charge the battery.
• Switches closed  grid is disconnected, so the inverter can perform propulsion
control of the machine.
26
Oscilloscope
displaying
waveforms
Laptop
DSP control
unit
Grid voltage
sensor
10 x 12V,
40Ah
LiFePO4
batteries
7.5kW DC/DC
converter and ninephase inverter
Nine-phase induction
machine
27
A small-scale controller looking at communication between EVs and the grid.
•
•
Display of real-time data for speed and
power flow.
•
•
Connect all vehicles to a network.
•
EV users notified about the demand on
the grid, and can export to help meet
demand, or charge up.
Shows whether demo supplies power to the grid
(V2G, power is -ve) or charges battery (G2V, power
is +ve).
Shows how fast battery is being dis/charged.
28

VESI project focuses on integrating the electrical
motor and power electronics

Meeting the key aims: reduce cost, increase power
density, improve reliability of electrical power
systems.

36 publications.

Underpinning basic research nearly complete.

Work on physical outputs for the 3 technology
demonstrators continuing.
29

VESI project information…….. l.murray@warwick.ac.uk

Semiconductors……………………..p.a.mawby@warwick.ac.uk

Demo 1/Motors ….……………….. p.c.k.luk@cranfield.ac.uk

Heat transfer….……………………..k.pullen@city.ac.uk

Power module cooling …………. volker.pickert@newcastle.ac.uk

Demo 2/Passives…………………… p.h.mellor@bristol.ac.uk
andrew.forsyth@manchester.ac.uk

Potting compounds ….………….. d.a.stone@sheffield.ac.uk

Packaging………………………………. lee.empringham@nottingham.ac.uk
mark.johnson@nottingham.ac.uk

Demo 3/On-board charging…. e.levi@ljmu.ac.uk

V2G functionality…………………..a.j.cruden@southampton.ac.uk
30