December 5-6, 2018 JMAG Users Conference 2018

Linear Motor & Eddy Current Brake Design For The Worlds First
Hyperloop Test Track

Arbi Gharakhani Siraki

EM Systems,
Virgin Hyperloop One

Abstract ：
Last year, the first full scale hyperloop test track, DevLoop, was successfully built and
tested in Apex, Nevada. By the end of 2017, a peak speed of 240mph (387kph) has been
reached on a track length of only 500 meters.
One of the many challenges was to design a linear motor that could take the pod to such a
speed in a short distance and later to design a fail-safe braking system, that brings the pod
to a complete stop on less than 200 meters of track.
This presentation will give insight into the hyperloop system and the design and analysis of
the DevLoop linear motor and the eddy current braking system.

24 - 1
December 5-6, 2018 JMAG Users Conference 2018

Linear Motor & Eddy Current Brake Design

For The Worlds First
Hyperloop Test Track

Dr Arbi G. Siraki
S r. M a n a g e r P ro p u l s i o n S y st e m s
1 © 2018 Virgin Hyperloop One Inc. All rights reserved. Confidential – Not for Distribution

OUR VISION

Fast, effortless journeys that expand

possibilities

24 - 2
December 5-6, 2018 JMAG Users Conference 2018

PA S S E N G E R | C A R G O

LOW-PRESSURE
TUBE
ELECTRO-MAGNETIC
PROPULSION

NEXT GEN MAGNETIC

L E V I TAT I O N

AUTONOMOUS CONTROL
P L AT F O R M

300 m/s
1080 km/h
670 mph

SPEED TRANSFORMS GEOGRAPHY

Virgin Hyperloop One – 30 min

V

Airline (with check-in) – 2.5+ hours

A

High-Speed Rail – 2+ hours

H

Road – 4+ hours
R

A B
500 km | 310 miles

6 © 2018 Virgin Hyperloop One Inc. All rights reserved.

24 - 3
December 5-6, 2018 JMAG Users Conference 2018

THE NEW PACKET NETWORK

Port
Inland Distribution

Commute
Zone

Hyperloop One – 0h 30m

A B
Regional
Airport
Centre

Suburban Area
500 km

THE NEW PACKET NETWORK

Port
Inland Distribution

Commute
Zone

Hyperloop One – 0h 19m

A B
Regional
Airport
Centre

Suburban Area
500 km

Radically Efficient

REDUCES CONGESTION &

EMISSIONS

100% ELECTRIC,
ZERO EMISSIONS

EFFICIENT LAND USE,

MINIMAL NOISE

24 - 4
December 5-6, 2018 JMAG Users Conference 2018

Making Hyperloop A Reality …Now

T ECHNOLOGY PEOPLE RESOURCES

XP-1
NORTH LAS VEGAS, NEVADA

First Hyperloop
One vehicle

DEVLOOP
NORTH LAS VEGAS, NEVADA

World’s first
full scale
Hyperloop
P H A S E 3 T EST S
DEC 14, 2017

387 km/h, 240 mph

24 - 5
December 5-6, 2018 JMAG Users Conference 2018

Linear Motor Design

For Acceleration of the DevLoop Pod

Desi gn Highlights
¾ Initial Target: Reach 300m/s within 2km distance (2.3g of acceleration)
¾ Designed for full system proof of concept not carrying passengers
¾ Motor design can be simply modified for lower accelerations ( around 0.25g ) for carrying passengers

¾ Wayside propulsion system

¾ Medium voltage motor
¾ Segmented track connected to MV VFDs through MV switches
¾ Low stator inductance a key design factor
¾ Low stator cost target
¾ Decoupled levitation and propulsion system, no motor force in levitation or guidance
directions

14 © 2018 Virgin Hyperloop One Inc. All rights reserved.

Devloop Longitudinal Layout

Propulsion and braking section

DevLoop Tube (500m)

300m 200m

Acceleration Deceleration
With LPMSM With Eddy current brake

0mph 240mph 0mph

(390km/h)

™ 390km/h in 300m is around 2g of acceleration

15 © 2018 Virgin Hyperloop One Inc. All rights reserved.

24 - 6
December 5-6, 2018 JMAG Users Conference 2018

System Archi tecture

Propulsion variable frequency drive Motor segment switch

• Multi-MVA VFDs • Custom design and integration
• Control integration with linear motor • Gen1 design tested at open air test

16 © 2018 Virgin Hyperloop One Inc. All rights reserved.

Motor Desi gn Choices

Stator Design
¾ Air core stator ( No Ferromagnetic core )
9 Significantly lower cost: No laminated electrical steel required for the track
9 No iron losses for high speed operation
9 Motor forces only in the propulsion direction
9 Lower motor inductance: longer motor sections, lower switch cost

Rotor Design
¾ PM synchronous motor topology was selected.
Double sided PM rotor was proposed for maximum flux generation.

¾ Following double sided PM architectures were studied

™ ŽƵďůĞƐŝĚĞĚ,ĂůďĂĐŚ ĂƌƌĂLJ͗Best option
™ ŽƵďůĞƐŝĚĞĚ,ĂůďĂĐŚ ĂƌƌĂLJǁŝƚŚďĂĐŬŝƌŽŶ: Lower force/mass ratio
™ ŽƵďůĞƐŝĚĞĚ,ĂůďĂĐŚ ĂƌƌĂLJǁŝƚŚŝŶƚĞƌƐƉĞƌƐĞĚŝƌŽŶƐĞŐŵĞŶƚƐ: Lower force/mass ratio & higher coil inductance
™ ŽŶƐĞƋƵĞŶƚŵĂŐŶĞƚĂƌƌĂLJǁŝƚŚďĂĐŬŝƌŽŶ : Similar force/mass ratio & higher coil inductance

¾ Embedded magnet options not considered due to large air gap and pole pitch.
17 © 2018 Virgin Hyperloop One Inc. All rights reserved.

Attractive Forces Between Magnet Arrays

Array to Array Attractive Force, Per Side Per Pole
Pair
16,000

14,000

12,000

10,000
Force [N]

8,000

6,000

4,000

2,000

0
50 100 150 200 250
Gap Between Arrays [mm]

Attractive load, 2D simulation Attractive load, 3D simulation

Measured

18 © 2018 Virgin Hyperloop One Inc. All rights reserved.

24 - 7
December 5-6, 2018 JMAG Users Conference 2018

Stator Design: Distributed Winding

Option.1: Q=2 Option. 2: Q=1

x Option 2 was selected. Rational:

¾ Lower total coil manufacturing
cost per block,
¾ Reduced mold complexity,
¾ Reduced number of busbars,
¾ Increased the spacing between
coils to have more insulation
¾ Better winding factor ሺ‫ܭ‬௪ ൌ ͳ.)
‫ܭ‬௪ ൌ ͲǤͻ͸͸ ‫ܭ‬௪ ൌ ͳ

19 © 2018 Virgin Hyperloop One Inc. All rights reserved.

Stator Design: Concentrated Winding

Motivation:
¾ Shorter end winding thus lower copper losses and lower end-
winding contribution to total inductance
¾ Ease of manufacturing

Design Spec:
¾ A design with 1 coil/pole/phase (q=0.5) is considered
¾ No subharmonics
¾ Theoretical winding factor of ‫ܭ‬௪ ൌ ͲǤͺ͸͸

20 © 2018 Virgin Hyperloop One Inc. All rights reserved.

Concentrated Winding: Winding Factor

The real winding factor can be found by comparing

the total generated force of concentrated winding
versus the distributed one ሺ‫ܭ‬௪ =1).

‫ܹ݃݊݅݀݊݅ ݀݁ݐܽݎݐ݁݊݁ܿ݊݋ܥ‬
‫ܭ‬௪ ൌ ൌ ͲǤ͸
‫ܹ݃݊݅݀݊݅ ݀݁ݐݑܾ݅ݐݏ݅ܦ‬

Lower winding factor is mainly due to large width

of each concentrated winding coils facing the
airgap.

21 © 2018 Virgin Hyperloop One Inc. All rights reserved.

24 - 8
December 5-6, 2018 JMAG Users Conference 2018

Eddy Current Loss in Magnets

¾ High order harmonics are moving at a different speed

compared to the magnets

¾ They are source of eddy current losses in magnets.

¾ The total loss per pole pair at full speed is estimated to be

around 24.4kW

¾ Considering thermal capacity of the magnets, 7.6W/kg/C,

and the volume, the temperature rise is estimated to be
around 39.6 C°/s which is not acceptable.
22 © 2018 Virgin Hyperloop One Inc. All rights reserved.

Eddy Current Brake Design

For Deceleration of the DevLoop Pod

Devloop Braking System

Requirements

• Deceleration from 115m/s to 0 in < 180m

-> 4G deceleration with ~50kN average braking force
• High reliability
• Low pod-side weight
• Minimal impact on pod dynamics
Potential options
• Regen with LPMSM
• Friction Brake
• Eddy current brake (separate from motor)
• Eddy current brake (using LPMSM magnets)
• …
24 © 2018 Virgin Hyperloop One Inc. All rights reserved.

24 - 9
December 5-6, 2018 JMAG Users Conference 2018

Material and –thickness Selection

• 3D transient magnetic FEA

• Performance for different
materials and thicknesses

• Aluminum
• Copper
• Ferromagnetic Steel
• Stainless Steel

25 © 2018 Virgin Hyperloop One Inc. All rights reserved.

Material and –thickness Selection

• 60kN peak braking forces

• 2-Zone braking region: start with
0.75”, then transition to 1.5” for
maximum braking
• Transition from .75” to 1.5” set at
100m to achieve minimum braking
distance (180m) for 115m/s initial
speed
• Braking distance from 100 m/s: 150m

26 © 2018 Virgin Hyperloop One Inc. All rights reserved.

Brake Rail design

Including transition regions

3m 90m 3m 80m

Entry Section High-Speed Section Ramp Section Low-Speed Section

• Fin is modified • constant fin thickness (19mm) • Width is varied to • constant fin thickness (38mm)
to reduce • Fin section lengths are sized at an odd reduce dynamic • Same fin spacing as high-speed
braking and multiple of the array pole pitch to reduce loads at 19- section
dynamic loads temporary loss of force at junctions 38mm transition • Nominally 60-0 m/s pod speed
on pod frame • Nominally 100-60 m/s pod velocity

27 © 2018 Virgin Hyperloop One Inc. All rights reserved.

24 - 10
December 5-6, 2018 JMAG Users Conference 2018

3D single pole (antiperiodic) model

Eddy Currents in Brackets

Brake rail
Magnets (onboard)
• Brackets attaching rail to tube are made
of ferromagnetic steel
Detailed skin depth mesh to
• Potentially high lift force capture eddy current path

• High skin effect in ferromagnetic steel

• 3D eddy current path
Brake Rail
Austenitic stainless steel

Brackets
Ferromagnetic carbon steel
Attaching rail to tube

28 © 2018 Virgin Hyperloop One Inc. All rights reserved.

Vector: Current Density

Contour: Magnetic Flux Density

Eddy Currents in Brackets

• Reduction of lift force by Eddy current in

80% by insulating brackets stainless steel
no skin effect
from fin 2D current path

Lift Force on Pod

Electrically connected to fin 1752N
Electrically insulated from fin 360N Eddy current in
ferromagnetic steel
strong skin effect,
Skin depth for ߤ௥ = 1000:
3-D current path
ʹߩ
ߜൌ ൌ ૙Ǥ ૝૟‫ܕܕ‬
ߤ߱

29 © 2018 Virgin Hyperloop One Inc. All rights reserved.

Test Results
Deceleration profile of high speed run

Final part of
acceleration

Transition from
acceleration to braking
1.5g -> -4g)
(1

30 © 2018 Virgin Hyperloop One Inc. All rights reserved.

24 - 11
December 5-6, 2018 JMAG Users Conference 2018

Test Results Brake Force vs Speed

(measured and predicted)

• Excellent agreement
between JMAG FEA and
measurement
• Error <3.3% across full
operating range

31 © 2018 Virgin Hyperloop One Inc. All rights reserved.

2018 Focus
S t ro n g G o ve r n m e n t - S p o n s o re d P ro j e c t s G l o b a l l y

UNITED ARAB
E M I R AT E S INDIA

USA

The 5th Mode of

Transportation is Closer
Than You Think

24 - 12
December 5-6, 2018 JMAG Users Conference 2018

Questions?

W W W. H Y P E R LO O P - O N E .CO M

@ H Y P E R LO O P O N E

24 - 13