Vehicle Design Summit

Electric Hub Motor (V2)

Eric Conner
Harvey Tang
Matthew Peddie
 Motivation
• The AHPV from VDS 1.0
used an expensive, NGM
electric hub motor,
costing roughly $8000.
(picture on right)

• VDS 1.0 required a new

electric hub motor to
serve as both a
replacement for the NGM
motor, and as a stepping
stone design for VDS 2.0.
 Requirements (Preliminary)

• 10 kW continuous Power
• 90%+ efficiency optimized for 45 miles an hour.
• Motor weight less than 30 kg
• Must interface with EV-C200 controller
• Acceleration from 0-60 mph in less than 15 s.
• Solar/Battery power must be used
• Constant Torque with speed variation
 Constraints
• Motor must fit between wheel
and suspension arm, not
interfere with other
components
• Motor cannot draw more
power then controller can
supply
• Torque must not
surpass limit of
suspension arm bolt
hole
 Preliminary Design Choices
• Design Choices Æ Why did we decide to design a 3-phase, axial
gap, double sided, slotted, surface mounted brushless DC motor?
Note, these design choices were made based on research, not
simulated optimization.
• Brushless Hub Motor Æ Comparison to Brush Hub Motor
– higher efficiency and reliability (reduction of electromagnetic
interference)
– reduced noise
– longer lifetime (no brush erosion)
– However, more difficult to control (resolved by digital control)
• Why 3-phase?
– Excellent starting conditions with smooth rotation and low
torque ripple Æ No structural resonance and induced
mechanical stress
– Flexible Æ Work with large variety of magnet configurations,
winding configurations, and coil winding
– Good conductor utilization Æ Higher phases give better
utilization but are offset by increased numbers of leads and
transistors
 Preliminary Design Choices (Continued)

• Why axial gap?

– Spatial limitations Æ Motor must interface with suspension arm;
fixed dimensions.
– Axial Gap gives compact machine construction and short frame
with much shorter rotor in axial direction, and thus less overall
thickness
– High power density.
– High efficiency; no rotor copper losses due to permanent-
magnet excitation.
• Why double sided air gap?
– The high attractive force between the rotor and the stator is
counterbalanced by the use of a second stator.
– Reduced copper and iron losses
– Increased power density.
– Increased cooling characteristics
 Preliminary Design Choices (Continued)

• Why slotted armature?

– A motor with armature slots is more robust
– Allowance for different winding structures
– Although the slotted armature implies increased
losses from flux ripple and tooth iron losses, the
increased robustness is necessary to combat the
mechanical stress.
– Slotted armatures give higher airgap flux density
levels using fewer permanent magnets.
• Why surface mounted permanent magnets?
– Much easier construction and manufacturing
compared to interior permanent magnets
 Design Variables
• How did we optimize our motors number of poles, stator slots,
magnet span, coil turns, magnet grade, magnet skew and air gap
length?
• Maxwell3D was used as a means of running dynamic
optimization Æ Program errors would not allow use of Maxwell’s
optimization toolbox Æ Several configurations were analyzed
separately so that
various trends
could be analyzed
for an optimized
engine

• On the right is the

sketch of the engine
prior to the
optimization of the
design variables
 Design Variables (Continued)
• The table on the right shows
Design Optimized
our final optimized engine and
Variable Value
the values for each design
variable Pole Number 8
Stator Slots 18
Magnet Grade NeFeB

Coil Span 2
Air Gap 1 mm
Stator Offset 15 Deg.
Magnet Span 150 Deg.
Magnet Skew 1 slot pitch
Wire Diameter .82 mm
 Design Variables (Continued)
• Pole Number Æ Smooth torque coupled with low speed generally
implies large pole count
• 8 poles decreases thickness of rotor yoke/stator yoke, decreasing
overall diameter.
• 8 poles minimizes flux leakage inside rotor
• 8 poles increases the axial length of the stator and the end
windings which reduces copper losses and increases efficiency
• Stator Slots Æ Related to pole number; slot/pole number must be
fraction to reduce cogging and skewing of poles or lamination stack.
• 18 slots gives coil span of 2 Æ easier to manufacture
• 18 slots reduces cogging torque
• 18 slots reduces the length of the end windings and consequently
the copper losses.
• Air Gap Length Æ Increased length results in more overall losses
while too small of a gap results in decreased power density
 Design Variables (Continued)
• Magnet Grade Æ NeFeB has a larger energy-density then other
magnets at a reasonable cost, increasing overall power density and
torque
• Stator offset Æ 15 degree offset of stators with each other was
arrived at; compromise between elimination of some higher order
harmonic components (decreases overall losses) and axial
asymmetry which can cause pulsating axial force and create losses.
• Magnet Skew Æ Skew can eliminate cogging torque as well as
high-frequency components related to flux losses
• Magnet Span Æ Span minimizes the pulsating torque, and in turn,
cogging torque.
• Wire Diameter Æ Optimized to turns per coil in the motor. Larger
diameter gives less losses, however, less turns per coil.
• Coil Span Æ Given by slot/poles, rounded down for short-pitching;
gives an increased machine efficiency by reducing the end-turn
lengths.
 Stator Slot Design
• Previously defined Maxwell3D slot configuration Section Size
for axial gap hub motors was used Wedge 1 mm
• Slot too deep or narrow Æ increased leakage Height
• Slot width too large Æ slot tooth saturation
Body 8 mm
• Slot top too open Æ cogging torque increases
Height
• Slot top too closed Æleakage will increase.
Opening 2.5 mm
Width
Wedge 6 mm
Max
Width
Bottom 6 mm
Width
Bottom 3 mm
Fillet
Opening 1 mm
Height
 Motor Geometry
Component Size (mm)
Inner 252 mm
Diameter
Outer 360 mm
Diameter
Rotor 36 mm
Thickness
Air Gap (x2) 1 mm
Stator 8 mm
Thickness (x2)
Frame 16 mm • Although inner and outer radius are good
design variables, in our case, we were
Thickness (x2) limited by the given AHPV dimensions.
Overall 86 mm We did, however, optimize the inner
Thickness diameter within the given constraints.
 Performance Analysis
• Maxwell3D ran simulations on various
inputs spanning several values for
each design variable Æ Hybrid
method of research and computer
aided analysis was used to select
final values. Below are torque and
speed graphs, and to the right is our
model representation within Maxwell
 Manufacturing
• Materials
• Copper – conductive
• Steel – cheap, strong; placed to minimize magnetic losses and
side effects
• Polycarbonate – strong, light, impact-resistant and easy to
machine
• Automotive bearings for thrust and radial support

• Techniques
• Milling
• polycarbonate shell
• smaller steel parts
• Waterjetting steel frame and rotor disc
• Epoxying
• magnets to rotor disc
• stator coils within polycarb shell
• Welding rotor disc and support discs to axle
 Assembly
• Overall Design Principles
• Simple
• Strong
• Light
• Rotor disc attached to axle
• Polycarbonate safety shell doubles as stator

structure
• Coils and power buss embedded
• Axle held in place by huge bearings
• Steel arms lock two halves of safety shell together

• Assembles rapidly
 Section Rate Qty Total Cost

Design and Engineering 0

Engineer $75* 40** 3000

CAD workstation $55 10 550

Simulation workstation

Rotor Disc
$55 30 1650
Cost
Magnets $30 32 960

Disc $35 1 35

Epoxy $10 1 10

Assembly

Driveshaft
$40 4 160
• Going for cheap
Steel Rod $33 1 33
• This table includes labor
Bearings $65 1 65

Steel Cylinders $15 2 30 • One-time costs of design

Bolts $15 1 15

Machining $50 2 100 • Non-bulk rates for materials

Assembly

Coils
$40 2 80
• Machine-shop rates for
Copper wire

Winding machine
$12

$75
36

3
432

225
machines
Lamination $40 4 160

Epoxy $10 3 30

Assembly $40 5 200

Shell

Polycarbonate $500 2 1000

Steel $150 1 150

Waterjetting $75 2 150

Other Machining $50 2 100

Assembly $40 4 160

Electronics

Hall Sensors $2 9 18

Wiring and regulation $40 1 40

Assembly $40 1 40

Total Cost 9393

 Future Work

¾1. Continuing Analysis Refinement

• Moving from parameters chosen originally by “rules of thumb” to CAD models
with finer detail resolution to capture these design decisions
• FEA simulations that span full dynamic response of the motor system including
startup acceleration and constant speed cruising

¾2. Motor Controller Design and Integration

• Developing power electronics capable of supplying the minimum 10kW with
sufficient scalability to accommodate future design demands
• Baseline software interface to ease monitoring of motor and wheel interaction
allowing programmable performance commands (ie. Anti-lock brakes)

¾3. Cost Reduction in Manufacturing Process

• Expand supplier base and list of coil winding companies
 Lessons Learned

¾ 1. Get expert advice from the start

• Reduces confusion about basic design differences and nomenclature
¾ 2. Get proper simulation software
• Having incompatible software made the optimization nearly possible
¾ 3. Constraint on overall design variables
• Motor design offers more variables than can be accounted for, reduction of
design variables preferable
¾ 4. Further research and time required for complex new design