Newcastle University ePrints - eprint.ncl.ac.uk

Yang S, Baker N, Mecrow BC, Hilton C,

Sooriyakumar G, Kostic-Perovic D, Fraser A.

Cost Reduction of a Permanent Magnet In Wheel Electric Vehicle Traction

Motor.

In: International Conference on Electrical Machines (ICEM 2014).

2-5 September 2014, Berlin: IEEE.

Copyright:

© 2014 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all

other uses, in any current or future media, including reprinting/republishing this material for advertising

or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or

reuse of any copyrighted component of this work in other works.

DOI link to paper:

http://dx.doi.org/10.1109/ICELMACH.2014.6960218

Date deposited:

01/10/2014

Cost Reduction Electric

Sichao Yang1, Nick J. Baker 1, B. C. M

1School of Electrical, El2Pro

Abstract -- This paper describes a motor fowheel of an electric vehicle. It demonstrates various rotor parameters on an outer rotor pemotor (ORPM) with Surface-mounted magnets this paper is to reduce the magnet volume, whiltorque performance through the complete opedifferent magnet topologies are investigated firSurface-mounted Permanent Magnet (SSPM), IPM (IPM) and V shape interior PM (VPM) desigin terms of torque capability at certain magnet vgives highest torque performance. The iron shiVPM can protect the magnets from the opposinthus providing increased resistance to demahence permitting thinner magnets. Furthermorcombination with higher torque capability hHowever, due to the increased inductance, modesign needs to work at a poorer power factor a reduced speed range for a given inverter. Lasand simple manufacture method of the VPaddressed with consideration of mechanical feas

Keywords – Cost Reduction, In-wheel Motorshape magnets

I. INTRODUCTION In recent decades, attention on Electric Ve

been greatly increased due to their high effemissions and reduced noise pollution. Howecost and a short driving range are two major electric propulsion system. The required htorque density leads the designer to use PeMachines, which along with the power electrothe batteries, produce a high initial cost.

The cost ratio between motor materials iCu : Si-Steel = 10 : 1 : 0.15 in [1].

Therefore, there is a drive to reduce the whilst keeping the motor cost effective, efficieThe motors in question, manufactured by Protmounted in the wheel of the vehicle ensuring toutput torque is available at the wheel, givibetter control and improving efficiency and eliminating parts of mechanical design, sucdifferentials, at the cost of increased size anmotor. In this project, the motor used as thmotor is a high-torque, low-speed, outer magnet motor with a good overload capabicontrol range, and designed to fit within aBasic performance parameters are shown in T

The aim of this project is to reduce the cost1.) using less magnet material; 2.) lowering th3.) simplifying assembly and manufacturin

of a Permanent Magnet Vehicle Traction Motor

Mecrow1, Chris Hilton2, Gunaratnam Sooriyakumar2,Al Fraser2

ectronic and Engineering, Newcastle University, Newcastle, United King

otean Electric Limited, Farnham, Surrey, United Kingdom

or use within the the influence of

ermanent magnet (SM). The aim of

le maintaining the erating range. Six rstly. Then, Semi I shape tangential gns are compared volume. The VPM ielding concept in ng armature flux, agnetisation, and re, a new slot/pole has been studied. tor with V shape and consequently

stly, Cost effective PM rotor is also sibility.

r, Outer Rotor, V

ehicles (EVs) has fficiency, lack of ever, high initial problems for the high power and rmanent Magnet onic inverter and

s given as PM :

magnet content, ent and compact. tean Electric, are that all the motor ing the customer compatibility by ch as gears and nd weight of the he benchmarking rotor permanent ility, wide speed a 16” wheel-rim. able I.

t of the motor by: he magnet grade; ng methods; 4.)

recycling scrap more efficiently. volume is the sole focus in this pape

II. BENCHMARK

Fig. 1. 2D model for one sub moto

Fig. 1 shows a 2D model of 1/6permanent magnet motor. This macthe benchmarking motor (BMMconcentrated double-layer windinground a single tooth, as illustratemotor parameters are listed in Table

Fig. 2. Winding configuration for

Table I Benchmarking

The BMM design was dridimensions and required torque peris constrained to 35mm, and configuration is selected to ensurwinding. As a given air-gap f

In-wheel r , D. Kostic-Perovic2 and

gdom

The reduction of magnet er.

KING MOTOR

or of the benchmark machine

6 of the existing outer rotor chine is here referred to as M). The motor has a g, with each coil wound

ed in Fig. 2. The primary e I:

r three phases of the motor

Motor Parameters

ven by standard wheel rformance. The axial length so concentrated winding

re a non-overlapping end-force produces a torque

proportional to its radius, to maximize torpossible air-gap radius is chosen. The ouconstrained by the wheel size, so the aimaximized by increasing the number of polethe required rotor core back depth. 48 poles wgood compromise. The motor is low speed dthe electric frequency at maximum speed (100iron loss in the motor is manageable and tswitching of the inverter does not create undue

A fractional number of slots per pole usnumber of slots for a given number of polesmanufacturing advantage and is also concogging torque. It also enables a significantachievable machine inductance to facilitate operation over a wide speed range with control, explained in detail in [3] and [4].

For fault tolerance, the machine can be splor six independent sub-motors. Multiphase dedecouple the flux from each phase with fault-this machine, multiphase sub-motors are arrangement offers sufficient fault tolerant sub motors fail. A general description of ttesting of the thermal, mechanical and independence of each sub is conducted in [8].

In order to be cost-competitive in the autthe volume of magnet, which is the most expethe motor, needs to be reduced. Therefore, topologies are studied to seek better magnecompared to this benchmarking motor

III. ROTOR TOPOLOGIES In the analysis of the benchmarking moto

conclusion is made by observing the impthickness on the magnet utilization: in esstorque capability per unit magnet mass. thickness is reduced, which can be seen in Fiutilization can be improved rapidly due toverall permeability in a less saturated magncircuit as indicated from Fig. 3 - also stated in

Fig. 3. Flux Linkage with Magnet Thickness chbenchmarking motor

Fig. 4. The Magnet Thickness change in the benc

rque, the largest uter diameter is ir-gap radius is e pairs to reduce

were found to be a direct-driven, and 00rpm) is 400Hz; the frequency of e switching loss.

sed as a smaller s gives a distinct nducive to low t increase in the

constant power flux weakening

lit into two, three esigns in [5, 6, 7] -tolerant teeth. In adopted. This capability when

the machine and electromagnetic

tomobile market, ensive material in

alternative rotor et utilization and

or, one important mpact of magnet

ence this is the As the magnet

ig. 4, the magnet to the improved etic flux flowing [9].

hanging on the

chmarking motor

Fig. 5. Torque Performance with Magbenchmarking m

By halving the magnet thickness4, the value of Torque / Magnet Mathe red curve in Fig. 5, is improvetorque at rated load is only reduced

The electric loading must be hidensity demand. The stator is covjacket, but it is still relatively diffinear the air gap periphery. The temexpected to be 100 degrees under cdemagnetisation is a potential issuevolume, as can be seen from Fig. 6flux density is above the magnet kneindicate where demagnetisation is occu

Fig. 6. Demagnetization analysis in toverload condi

The benchmarking motor is 90Amps peak over-loading currentmagnets are narrowed demagnetismagnet topologies are now invesolutions.

In general there are three categorotor: surface-mounted magnet, suinterior magnet. According to [10]has better demagnetisation resistandifferent magnet positions were mterms of peak torque performaresistance at the rated condition, as

gnet Thickness changing on the motor

s to 2mm, as shown in Fig. ass, which is illustrated with ed by 65% whilst the peak by 18%.

igh due to the high torque vered with a fluid cooling cult to remove rotor losses

mperature of the magnets is continuous running. Hence, e when reducing the magnet . The blue regions indicate the ee-point and the red regions

urring.

the benchmarking motor under tions

not demagnetized with t input, but as soon as the sation occurs. So the new estigated to seek possible

ories of permanent magnet urface-inserted magnet and ], the interior magnet rotor nce. Six new models with modelled and compared in ance and demagnetisation shown in Fig. 7:

Fig. 7. Cross sections of: a.) Triangle PM; b.) Surfaceshaped PM; d.) V-shaped PM; e.) Spoke type PM;

In this study the Stator design is unalterrotor depth is fixed at 9 mm, which only leback in the Surface-mounted Permanent MThe flux is highly saturated in the Buried Pe(BPM) design and torque performance is consThe Triangle shape is developed based on the[11]. After its primary optimisation, the peaksimilar to the benchmarking motor, but usmaterial. The idea of a C shape design, which[12], is similar to the V shape: it produces fland increased inductance. Consequently, the Cand spoke shape designs give a higher tocomparison with the benchmarking motor. HoPM and spoke shape PMs were selected for fto their relatively simple manufacturing proce

IV. V SHAPE DESIGN In Fig. 8, the best performance is seen

design. The blue line stands for the benchwhose upper bound of magnet mass is itsCorrespondingly, the magnet mass is reduthinning the magnets at its lower bound. Simivolume is reduced by changing the thickneSpoke designs as shown in Fig. 9 and Fig. 10.

The V shape can give the same torque wmagnet mass, whilst the spoke shape needs tooriginal material to generate the same torque.

Fig. 8. Torque Performance with Magnet Thicknesbenchmarking motor, V shape PM and Spoke

e-mounted PM; c.) C-

f.) Buried PM

red and the total eaves 5 mm core Magnet (SMPM). ermanent Magnet sequently poorer. e motor shown in k torque value is ing less magnet h is expressed in lux concentration C shape, V shape orque density in owever, V shape further study due ss.

in the V shape hmarking motor, s original value. uced to half by larly, the magnet ss in the V and

with 56% of the o use 78% of the

ss changing on the shape PM

Fig. 9. The Magnet Thickness

Fig. 10. The Magnet Thickness c

As for the demagnetization issuno demagnetisation under the sameshown in Fig. 11. This is because thmagnets acts as a protecting shiefrom high armature magnetising opening, as seen in Fig. 12.

Fig. 11. Demagnetization

Fig. 12. Field plot with Flux flowin

The variation of spoke shape performance improvement has beendevelopment is focused on the V sh

Fig. 13 and Fig. 14 show theoriginal, surface mounted, benchmshaped design at rated condition. Ton full current advance angle rangpermeability method.

s change in V shape PM

change in Spoke shape PM

ue: the V shape design has e over-loading condition, as he iron piece in front of the eld, protecting the magnet

field strength at the slot

n analysis in VPM

g direction in V shape design

has been studied and no n found. Therefore, further

hape design.

e torque capability of the marking design, and the V The separate torque curves ge are generated by frozen

Fig. 13. Torque vs. Current Advance Angle curve inmotor

Fig. 14. Torque vs. Current Advance Angle cur

Due to the contribution of the reluctance shape design, the magnet volume can be reduwhilst maintaining the torque output. The comtorque values in different designs is given benchmarking motor with 56% of the magnthe lowest torque output, whereas the V shapemagnet volume has 97% peak torque cobenchmarking motor.

Therefore, the V shaped design with performance compared to BMM is chosen toconcern of mass production.

Fig. 15. Torque Performance Comparison on

V. V SHAPE MODIFICATION

A. Pole / Slot number investigation The level of flux concentration within the

is a function of the magnet width compared tover which it spans. For a fixed rotor radial reduces with increasing pole number. In ordthe impact on performance for a given magnnumbers have been investigated. As before, remains unchanged with 54 slots. By incrnumber from 48 to 60 and retaining ththickness and its overall mass, the design chan

n the benchmarking

rve in V shape

torque in the V uced significantly mparison of peak in Fig. 15. The

net volume gives ed design with 56% ompared to the

matched torque o be modified in

BMM and V

N

V shaped design to the air-gap arc

depth, this ratio der to understand et, different pole the stator design reasing the pole e same magnet nges as shown in

Fig. 16. The 60 pole design has a higiven stator current.

Fig. 16. Rotor Topology in

Fig. 1 Torque Performance C

Compared to the BMM, torque the new design. However, the incrdemanding at the same mechanical and q axis inductance. In turn, thspeed for the machine. This situatithe higher pole number design. envelope is illustrated in Fig. 17:

Fig. 17. T vs. Speed of the benchma

The base speed of the V shap343rpm, compared to 482rpm in the48 poles, which means the 6% hobtained with the cost of losing 25%This could only be overcome by inthe inverter.

However, when the same test isdesign with 48 poles, the base speThe VA rating can thus be retrievedtorque in comparison to the higher p

Then, the number of coil turnspoles design and V60 poles desigtorque in BMM. With the same ra

igher torque capability for a

two V shape Designs

Comparison on V design

capability is maintained in reased electrical frequency speed results in increased d

his results in a lower base on is particularly severe in The torque against speed

arking motor and Air V shape

pe design with 60 poles is e benchmarking motor with higher torque capability is % of its rated output power. ncreasing the VA rating of

performed on the V shape eed is extended to 452rpm. d at the cost of slightly less pole number design.

s is changed in both V48 gn to match with the rated ated current value, the base

speed of BMM, V48p and V60p are 529rp415rpm, respectively.

The 48 poles design is remained due tpower compared to the 60 poles design.

B. Loss analysis

Fig. 2 efficiency on full speed range

As can be seen from Fig. 2, the efficiespeed is increasing above base speed point. Trelatively higher current advance angle for control in V shape and its high content of harm

The study to reduce the loss at high speed amaterial cost by replacing the back laminatiostack is still under study at the time of writing

C. Manufacturing consideration In order to ease the manufacturing process

assembly speed, it is preferred mechanicallamination to be a single piece, as shown in Fi

Fig. 18. Conventional V shape

However, this arrangement creates two flux either end of the magnet, which reduce the torq

Several designs have been created and anthe flux leakage path, as shown in Fig. 19:

Fig. 19. V shape modifications

pm, 469rpm and

to higher output

ncy drops while This is due to the

field weakening monics.

and to reduce the n with solid iron this paper.

and increase the lly for the rotor ig. 18:

leakage paths at que by 20%.

nalysed to block

The ideal V shape design is dmade rectangular in design 2 to rease the magnet cutting process. indicate the flux leakage paths, whi3. Now the single piece laminationwhich is difficult to assemble ontomechanical feature is introduced iinner rotor piece, hold outer rotor the rotor pieces. However, the key rflux, which reduces the flux linkage

Hence, the conventional designoff the inner flux flowing path in drotor back lamination in one piece.

In design 7, the outer flux flointroducing an air bridge, which reouter joint of the magnets. Also, ththe inner joint of the two magnets that area. Moreover, there is only sitting on the periphery of the innerthe demagnetization resistance preventing the magnet from direcflux travelled across the air-gap.

The field view is shown in performance in these designs is shoperformance of design 7 gives the(the ideal design) whilst making msimple due to a single core back lam

Fig. 20. The field view

Fig. 21. Torque Perfo

In summary, the peak torque vahas deteriorated by using a conven

design 1. The magnets are reduce material waste and However, the red circles

ich are eliminated in design n is made into multi-pieces, o the rim. The key shaped in design 4 to support the pieces together and locate

results in additional leakage e.

n 5 is improved by cutting design 6, while keeping the

owing path is blocked by educes flux leakage on the here is no leakage path on because of lack of iron in one point on each magnet

r rotor radius, which means is further increased by

ctly opposing the armature

Fig. 20 and the torque own in Fig. 21 . The torque e closest result to design 1 manufacture and assembly

mination.

w of Air Bridge V

ormance Chart

alue of the V shape design ntional way to gain an easy

manufacturing process. Therefore, a new air-created in the V shape which maintains performance with a single outer rotor piece. shape includes an air bridge on two adjacentform one pole.

D. Mechanical feasibility In the rotor there are two forces on the roto

the centrifugal force due to the rotation anattractive force from the adjacent magnets andacross the air gap. The centrifugal force is nproblem in this application, as the outer rim ctogether. The magnetic force has been simuby winding one coil around the closest tooth frotor piece and injecting the maximum currenin Fig. 3.

Fig. 3 The magnet attractive force test on the in

The worst scenario for force is illustrated the peak value of the overload phase currenopposite polarity of the magnets, which attrpiece, at a standstill. Together with Fig. 5, band the value of the forces are presented withstand this amount of force, the rotor piewould be glued onto the rotor back.

Fig. 4 The field view of the flux trave

Fig. 5 the direction of the force on each

Table 2 the exerted force on each piece (‘–‘ sign methe force is opposite to the reference direc

With further optimisation process, it wasrotor core back depth could be reduced, resultireduction in magnet volume, but only 4% red

number 1 2force perpendicular 67.37 2.01

tangential 1.79 -22.17

-bridge design is its peak torque This new design t magnets which

or pieces, namely nd the magnetic d the facing teeth not considered a

can hold the rotor lated in MagNet facing the testing nt, as it is shown

nner rotor piece

in Fig. 4, where nt (90A) has the racts the triangle both the direction

in Table 2. To eces and magnets

el path

h piece

eans the direction of ction)

s found that the ing in 7% further duction in torque

capability. This is shown in Fig. 22shown in Table III:

Fig. 22. Rotor Outer Radius cha

Table III Parameters of Air V a

Thus, the manufacturing cost issteel volume. Meanwhile, other pmaterial grade and changing rotor split ratio are under study.

VI. CONCL

With careful choice, it is possibmagnet volume employed in surfacwithout major loss of performandemagnetization. A surface mounmotor is replaced with V shape mmagnet mass. The demagnetization detailed modifications are madeproduction. The V shape design inductance, which must be clinductance becomes too large threduction in the torque – speed enve

VII. ACKNOWLEThe authors acknowledge the c

J. Ifedi for his work on the Application.

VIII. REFERE[1] C. Cheol Min, S. Jung H

"Comparative study of EVclass vehicles," in Electri(ICEMS), 2011 Internatiopp. 1-3.

[2] T. Finken, M. Hombitzerand comparison of severalrotor types regarding thevehicles," in Emobility -2010, 2010, pp. 1-7.

[3] D. Evans, Z. Azar, L. "Comparison of optimal dPM machines having nondifferent rotor topologiesMachines and Drives International Conference o

31.93 Nm-19.68

. The simulation results are

ange: 170mm to 166.5mm

and benchmarking motor

s further reduced with less otential aspects like lower diameter to stator diameter

LUSION ble to greatly decrease the ce mounted magnet rotors, nce or increased risk of

nted magnet benchmarking magnets, saving 44% of the

resistance is improved and e to accommodate mass

results in an increase in osely monitored. If the

hen there is a significant elope of the machine.

EDGMENT ontributions of Chukwuma previous version of this

ENCES Hoon, and N. Kwang-Hee, V propulsion motors for A-cal Machines and Systems

onal Conference on, 2011,

, and K. Hameyer, "Study l permanent-magnet excited eir applicability in electric - Electrical Power Train,

J. Wu, and Z. Q. Zhu, design and performance of -overlapping windings and s," in Power Electronics, (PEMD 2010), 5th IET on, 2010, pp. 1-7.

[4] W. Jiabin, K. Atallah, Z. Q. Zhu, and D. Howe, "Modular Three-Phase Permanent-Magnet Brushless Machines for In-Wheel Applications," Vehicular Technology, IEEE Transactions on, vol. 57, pp. 2714-2720, 2008.

[5] Q. Chen, G. Liu, L. Sun, Y. Jiang, and J. Yang, "Comparison of five topologies rotor permanent magnet motors with improved fault-tolerance," in Industrial Electronics (ISIE), 2013 IEEE International Symposium on, 2013, pp. 1-5.

[6] C. Qian, L. Guohai, Y. Junqin, G. Wensheng, and Z. Wenxiang, "Comparison Of Two Interior Permanent-Magnet Motors With Improved Fault-Tolerance," in IECON 2012 - 38th Annual Conference on IEEE Industrial Electronics Society, 2012, pp. 4093-4098.

[7] L. Guohai, G. Wensheng, C. Qian, J. Linni, S. Yue, and Z. Wenxiang, "Design and analysis of new fault-tolerant permanent magnet motors for four-wheel-driving electric vehicles," Journal of Applied Physics, vol. 111, pp. 07E713-07E713-3, 2012.

[8] C. J. Ifedi, B. C. Mecrow, S. T. M. Brockway, G. S. Boast, G. J. Atkinson, and D. Kostic-Perovic, "Fault-Tolerant In-Wheel Motor Topologies for High-Performance Electric Vehicles," Industry Applications, IEEE Transactions on, vol. 49, pp. 1249-1257, 2013.

[9] K. Mi-Jung, C. Su-Yeon, L. Ki-Doek, L. Jae-Jun, H. Jung-Ho, J. Tae-Chul, et al., "Torque Density Elevation in Concentrated Winding Interior PM Synchronous Motor With Minimized Magnet Volume," Magnetics, IEEE Transactions on, vol. 49, pp. 3334-3337, 2013.

[10] K. Ki-Chan, K. Kwangsoo, K. Hee-Jun, and L. Ju, "Demagnetization Analysis of Permanent Magnets According to Rotor Types of Interior Permanent Magnet Synchronous Motor," Magnetics, IEEE Transactions on, vol. 45, pp. 2799-2802, 2009.

[11] K. Yamazaki, M. Kumagai, T. Ikemi, and S. Ohki, "A Novel Rotor Design of Interior Permanent Magnet Synchronous Motors to Cope with Both Maximum Torque and Iron Loss Reduction," Industry Applications, IEEE Transactions on, vol. PP, pp. 1-1, 2013.

[12] K.-j. Lee, K. Kim, S. Kim, J.-S. Ahn, S. Lim, and L. Ju, "Optimal magnet shape to improve torque characteristics of interior permanent magnet synchronous motor," Journal of Applied Physics, vol. 97, pp. 10Q505-10Q505-3, 2005.

IX. BIOGRAPHIES Sichao Yang received a BE degree in Electric and its Automation Control

from South West University for Nationalities, China, in 2011, and a MSc degree in Electrical Power from Newcastle University, UK, in 2012. He is currently working towards the PhD degree in Newcastle University, designing a high torque, fault tolerant and cost competitive In-wheel motor for Electric Vehicles, which is sponsored by Protean Electric.

Nick J. Baker received a MEng Degree in Mechanical Engineering from Birmingham University, UK, in 1999 and a PhD from Durham University, UK, in 2003 for work in electrical machine design for marine renewable energy devices. He subsequently worked as an academic at Lancaster University (2005-2008), a renewable energy consultant at TNEI and presently

a Lecturer at Newcastle University’s Power Electronics Machines and Drives Group. Nick is a machine designer with research projects across the automotive, aerospace and renewable energy sector.

Barrie C. Mecrow is Professor of Electrical Power Engineering and head of the School of Electrical and Electronic Engineering at Newcastle University, UK. His research interests include fault tolerant drives, high performance PM machines and novel switched reluctance drives. He is actively involved with industry in the aerospace, automotive and consumer product sectors, who fund a large range of projects. Barrie commenced his career as a turbo-generator design engineer with NEI Parsons, England. He became a lecturer at the University of Newcastle in 1987 and a professor in 1998.

Chris Hilton is the Chief Technology Officer at Protean Electric with particular responsibility for advanced research, systems design, systems engineering and intellectual property. He has previously held roles in the fields of communications electronics, satellite navigation and particle physics research. Chris holds a PhD in physics from the University of Manchester, UK, and a first class honours degree in mathematics from the University of Cambridge, UK.

Gunaratnam Sooriyakumar is a Senior Development Engineer at Protean Electric Limited, UK where he is working on development of electric motor drives for automotive application. He obtained BSc in Electrical and Electronic Engineering from University of Peradeniya, Sri Lanka. He obtained his PhD from UEL with the sponsorship from Emerson industrial automation. He worked for Emerson industrial automation where he was the leader for R&D team which has developed the commercially successful next generation electric motors with higher torque density and high dynamic capability for industrial automation application. In addition he provided his expertise to improve the bespoke products at Emerson industrial automation which includes various kinds of electric motors for military and aerospace application. After leaving Emerson Industrial Automation, he continued working as an engineering consultant for Emerson industrial automation at Andover and for a wind generator design company at Edinburgh until joining Protean Electric Limited.

Dragica Kostic Perovic gained her first degree in electrical engineering at the University of Belgrade, Serbia, and her DPhil at the University of Sussex, UK. She is a Principal Motor Design Engineer at Protean Electric with main interests in the area of electromagnetic motor design, and DFMEA as a design process.

Alexander Fraser worked as a Senior Mechanical Systems Engineer at Protean Electric with a focus on motor conceptual design and dealing with vehicle-level engineering solutions related to the integration of in-wheel motors onto vehicle platforms. He holds a BEng Honours Degree in Automotive Engineering from Oxford Brookes University, UK.