ieee transactions on transportation …...ieee transactions on transportation electrification, vol....

IEEE TRANSACTIONS ON TRANSPORTATION ELECTRIFICATION, VOL. 2, NO. 3, SEPTEMBER 2016 391

High-Speed Solid Rotor Permanent MagnetMachines: Concept and Design

Puvan Arumugam, Member, IEEE, Zeyuan Xu, Antonino La Rocca, Gaurang Vakil, Matthew Dickinson,Emmanuel Amankwah, Tahar Hamiti, Serhiy Bozhko, Member, IEEE, Chris Gerada, Member, IEEE,

and Stephen J. Pickering

Abstract— This paper proposes a novel solid rotor topologyfor an interior permanent magnet (IPM) machine, adoptedin this case for an aircraft starter-generator design. The keychallenge in the design is to satisfy two operating conditionsthat are a high torque at start and a high speed at cruise.Conventional IPM topologies that are highly capable of extendedfield weakening are found to be limited at high speed due tostructural constraints associated with the rotor material. To adoptthe IPM concept for high-speed operation, it is proposed to adopta rotor constructed from semimagnetic stainless steel, which hasa higher yield strength than laminated silicon steel. To maintainminimal stress levels and also minimize the resultant eddy currentlosses due to the lack of laminations, different approaches areconsidered and studied. Finally, to achieve a better tradeoffbetween the structural and electromagnetic constraints, a novelslitted approach is implemented on the rotor. The proposedrotor topology is verified using electromagnetic, static structural,and dynamic structural finite-element analyses. An experimentis performed to confirm the feasibility of the proposed rotor.It is shown that the proposed solid rotor concept for an IPMfulfils the design requirements while satisfying the structural,thermal, and magnetic limitations.

Index Terms— Aircraft, high-speed machines, interior, moreelectric, permanent magnet (PM), solid rotor, starter-generator,structural.

I. INTRODUCTION

THE aircraft electrical power systems in the next genera-tion commercial aircraft are undergoing significant devel-

opment. Future aircraft power systems are expected to be morefuel efficient and also simpler to service and maintain. The waytoward this goal has been identified as a move toward moreelectric systems by the replacement of hydraulic and pneu-matic sources of power with electrical counterparts [1], [2].This can lead to an increased reliance on electrical powerfor a range of primary functions including actuation, de-icing,

Manuscript received March 10, 2016; revised May 29, 2016; acceptedJuly 4, 2016. Date of publication July 18, 2016; date of current versionAugust 16, 2016. This work was supported by the Seventh FrameworkProgramme within the Joint Technology Initiatives-Theme: Clean Sky throughthe Alternator with Active Power Rectification and Health MonitoringProject under Grant JTI-CS-2011-1-ECO-02-009. (Corresponding author:Puvan Arumugam.)

P. Arumugam and M. Dickinson are with Force Engineering Ltd.,Shepshed LE12 9NH, U.K. (e-mail: [email protected]; [email protected]).

T. Hamiti is with the Institut du Véhicule Décarboné Communicant et saMobilité, Versailles 78000, France (e-mail: [email protected]).

Z. Xu, A. La Rocca, G. Vakil, E. Amankwah, S. Bozhko, C. Gerada,and S. J. Pickering are with the Power Electronics Machines andControl Group, Faculty of Engineering, The University of Nottingham,Nottingham NG7 2RD, U.K. (e-mail: [email protected];[email protected]; [email protected]; [email protected]; [email protected]).

Digital Object Identifier 10.1109/TTE.2016.2592684

cabin air conditioning, and engine start. A more electric powergeneration system plays a key role in this technology, and thispaper focuses on the design of a starter-generator for suchsystems. One of the challenges often encountered in the designof a starter-generator for aero-engines is the need to satisfy twofundamental functions, namely, to energize the engines duringstartup and to generate power during normal engine operation.In addition, it should provide high power density [3]–[5] andsufficient reliability [6], [7].

An interior permanent magnet (IPM) machine with a wideconstant power to speed ratio is considered to be ideal for suchapplications [8]–[12]. In addition to the high energy efficiency,the IPM offers a high power to volume ratio [8], [9], [13]–[17].However, the operational speed of the IPM topology is signifi-cantly limited by the material yield strength due to centrifugalforce acting on the rotor body. The amount of force generatedin the rotor depends on the body mass and the operating speed.Since high-speed operation is required by the application, theremaining way to minimize this force is through downsizingthe rotor mass. This, however, will limit the electromagneticperformance of the machine as it reduces either the machine’slength or the rotor diameter, or both.

In [18], high silicon content laminated steel that has a con-siderably higher yield strength than the other laminated steel(e.g., nonoriented silicon steels: M270-50A and M250-35A)was employed in the design to improve the operational speed.Although an improvement can be achieved, the operationalspeed of the machine is limited by material yield strengthagainst the centrifugal forces acting on the machine. In [19],the design process has been done in such a way to finda fine balance between magnetic and structural limitations.To achieve both the required structural and magnetic per-formance, the power to volume ratio needs to be compro-mised. Recent studies show that the centrifugal force canbe used to extend the field weakening capability by addinga spring loaded mechanism with flux barriers [20], [21],introducing a self-activated flux weakening device/floatingparticles [22], or varying the permanent magnet (PM) direction[23]. Although these can be solutions to extend the fieldweakening, this would not improve the power density or thestructural capability of the machine. A way of improving thepower density of PM machines for high-speed applicationswhile also maintaining the structural integrity is subject toinvestigation.

The main objective of this paper is to present a novel con-cept for an IPM machine to overcome the electromagnetic and

2332-7782 © 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

392 IEEE TRANSACTIONS ON TRANSPORTATION ELECTRIFICATION, VOL. 2, NO. 3, SEPTEMBER 2016

Fig. 1. Required torque-speed characteristic of the aircraft starter-generator.

structural constraints of the target application. The applicationrequirements and the initial design are presented in Section II.Section III presents the proposed solid rotor solution for anIPM machine. To improve structural integrity and minimizeeddy current losses, three different rotor design approachesare considered, namely, a combination of both laminationsand solid steel, segmented solid steel, and slitted solid steel.A tradeoff study is carried out in Section IV. Section Vpresents a detailed structural analysis of an optimized machineto limit the localized stress acting on the rotor body. Herein, anovel design approach is proposed to overcome the structurallimitation while also compromising efficiency against the start-ing requirements. In Section VI, an experiment is performedto show the eddy current minimization through the proposedslitted approach in comparison with a solid rotor body. Thechallenges and limitations involved in the design and methodsto overcome such issues are discussed.

II. INITIAL DESIGN

The target application is to design a machine that acts as astarter and a generator, which feeds into a 270 V dc supplywith a nominal power of 45 kW at 32 kr/min. Fig. 1 showsthe required torque-speed characteristic. The machine runs asa motor during engine start and must supply constant torquefrom standstill to ωstart . Between ωstart and ωmin, the machineprovides a constant power to accelerate the engine. When themachine reaches ωmin, it operates as a generator that generatesa maximum power of 45 kW at a maximum speed (ωmax) of32 kr/min.

In the preliminary design, an IPM machine is designedto satisfy the electromagnetic requirements. To overcome thestructural limitations, high silicon content (6.5%) electricalsteel that has better mechanical properties and relatively poormagnetic properties has initially been adopted to allow amaximum stress of ∼800 MPa. Nevertheless, it has been notedthat although the designed machine is capable of extendedfield weakening, the stress limit still constitutes a bottleneckin reaching the maximum required speed of 32 kr/min. Only80% of the required maximum speed can be achieved whilefulfilling the peak torque requirement.

III. SOLID ROTOR IPM MACHINE

To overcome the structural limitations mentioned above andthus extend the operational speed required by the application,semimagnetic stainless steel that has high yield strength limit

Fig. 2. IPM rotor with (a) unmodified solid rotor, (b) axially segmentedrotor settled on stainless steel shaft, and (c) circumferentially slitted/milledsolid rotor.

and nonlinear isotropic magnetic permeability (maximum fluxdensity is ∼1.4T) is therefore proposed. The advantages ofhaving semimagnetic stainless steel over laminated magneticsteel are as follows:

1) one solid body that has higher yield strength and higherstiffness (better rotor dynamic performance);

2) better thermal properties (lower thermal resistance in theaxial direction);

3) semimagnetic; however, the bridges can easily be satu-rated.

The key disadvantage of semimagnetic stainless steel isthat the material has a relatively high electric conductivity,and therefore it is subject to higher eddy current losseswhen compared with laminated silicon steel. This could besignificantly minimized via a different structural adaptation.Within this paper, feasible structural adaptations are thereforeconsidered. Two different rotor arrangements are implementedin place of a complete solid rotor body as shown Fig. 2(a).Those are as follows:

1) segmented semimagnetic stainless steel rotor yoke fittedonto a nonmagnetic stainless steel shaft (Fig. 2(b)];

2) circumferentially slitted/milled (around the outer rotorperiphery) solid rotor body as shown in Fig. 2(c).

ARUMUGAM et al.: HIGH-SPEED SOLID ROTOR PM MACHINES 393

The analysis of the abovementioned arrangements is pre-sented via a tradeoff study in Section IV.

IV. TRADEOFF STUDY

By selecting a high-strength and high-stiffness materialfor the rotor, the limitation on the speed can be improvedsignificantly. As previously mentioned, the main disadvantageof a solid rotor is the excessive eddy current losses.In order to minimize these losses, the rotor can be made ofeither segmented solid disks or a slitted single rotor body[see Fig. 2(c)]. Instead of using solid conductive iron for thesegmented case, combinations of both high yield strength(∼800 MPa) silicon steel laminations and semimagneticstainless steel [see Fig. 2(b)] as a hybrid concept couldalso be considered. In such an arrangement, it is expectedthat the solid steel delivers the structural capability towithstand the centrifugal force at high speed, while thelaminated steel lowers the losses. In this section, thestructural limitations of such arrangements (the hybrid andthe slitted/milled rotor concept) are investigated. Throughoutthe analysis, the axial length of the rotor is kept the same. Themagnets are considered as a segmented piece with respectto the arrangement of the rotor. Stresses in the magnet areconstrained to 90% of the maximum allowable stress, andthus, cracks within the magnets can be avoided.

Considering the fact that laminated steel reduces the losseseffectively, an initial study is carried out for the combinationof laminated steel and solid steel. In the analysis, a number oflaminated steel and solid semimagnetic stainless steel cases,which have a thickness of x and y mm, respectively, andoccupy the complete stack length of the rotor, are considered.A solid thickness of 1, 2, 5, and 10 mm and an associated ratioof laminated steel from 50% to 550% with respect to solid steelare considered. The analysis is carried out using structuralfinite element (FE) where an importance is given to theimprovement on stress distribution within the entire structureby quantifying optimal thicknesses. The obtained maximumstress against the percentage variation of the laminated steelstack (x) with respect to solid steel (y) for different thicknessesof solid steel [see Fig. 3(a)] is presented in Fig. 4.

Fig. 4 shows the obtained maximum stress distributionfor different thicknesses of laminated steel with respect tothe thickness of solid steel. For example, if the solid steelthickness considered is 1 mm, the 250% of the laminatedsteel is 2.5 mm. The results indicate how the stress variesin accordance with the selection of the material thickness andthe optimum thickness ratio between solid steel to laminatedsteel, which provides minimal stress.

From the minimum stress line in Fig. 4, it can be seenthat the stresses vary with the solid disk thickness. It can alsobe seen that the stress fluctuates as the thickness changes.This is mainly because of the maximum stress that occurs atdifferent locations within the rotor with respect to changesin the material thickness. Between 50% and 300%, the stressreduction is significant and the most effective stress reductionfor all the stack combinations (x + y) occurs at 1:2 ratios(150%) of x and y. However, the maximum stress is still

Fig. 3. Arrangement of the rotor with different segmentations. (a) Combi-nation of laminated steel and solid steel. (b) Milled/slitted rotor with a space.

Fig. 4. Maximum stress occurring in laminated steel against percentagevariation of the laminated steel stack (x) with respect to solid steel (y) fordifferent thicknesses of solid steel.

higher than the yield strength of laminated steel, and thus, thisarrangement is not considered for further analysis. Althoughthe segmented solid steel can be used for the rotor, it wouldnot be beneficial compared with the milled/slitted arrangement.This is due to the single rotor body of the slitted arrangementthat allows for increased stiffness of the shaft when comparedwith the segmented case. The milled single rotor body [seeFig. 3(b)] is therefore analyzed with respect to optimizing thestack thickness and space in between.

Fig. 5 shows the obtained maximum stress for differentthicknesses of the slit space and associated slit thickness(1, 2, 5, 10, 15, and 20 mm). As can be seen fromFig. 5, the localized maximum stress in the rotor fluctuateswith both the thickness of the steel and the space betweenthem. It can also be seen that the obtained stresses for differentstack thickness are lower at three regions. Those minimalstress regions are M1 (0.2–0.4 mm), M2 (0.65–0.9 mm), andM3 (2.4–2.6 mm). Region M3 has a comparatively largerdistance between the stacks, but still shows low stresses.Region M1 is the optimal space distance for all the cases;however, this would be complex to produce due to the size ofthe wire and a limitation on manufacturing tools. Thus, a stack


Fig. 5. Maximum stress induced in solid steel against the space betweentwo stacks of solid steel (M1–M3 are the regions of minima).

Fig. 6. Structural geometry adaptations considered in the optimization.(a) Initially rotor design, (b)–(d) alternative rotor design.

thickness between 2 and 4 mm and a space distance rangefrom 0.2 to 0.6 mm are considered for further designoptimization.

V. ROTOR DESIGN AND ANALYSIS

As previously mentioned, a high strength stainlesssteel (PH17-4) that has a tensile strength of 1448 MPa has beenselected for the analysis. The active section of the rotor bodyis slitted or segmented so as to minimize eddy current losses.The magnet segments are then placed into the active sectionof the rotor body. To achieve better stress distribution on therotor while satisfying the torque-speed requirement shown inFig. 1, the areas associated with the magnet, hollows, wedges,and iron bridges [shown in Fig. 6(a)] are optimized. Bothelectromagnetic and structural optimizations are performedusing FE analysis.

A. Design Optimization

The structural optimization of the rotor begins with aninitial IPM rotor design based on the FE magnetic analysis.In the initial design, an analytical model [18] estimates thestresses due to the centrifugal forces that act on the bridges,by representing a ring equivalent to the mass of both theiron bridge and the magnet. This is used along with FEelectromagnetic analysis. The rotor is optimized consideringthe torque-speed requirement, and then subject to a detailed

TABLE I

STRESS DISTRIBUTION FOR DIFFERENT STRUCTURAL ADAPTATIONS

Fig. 7. Stress distribution (MPa) in the IPM rotor. (a) Initial design (case A).(b) Optimized first set design (case B3). (Units are in megapascals.)

coupled structural and magnetic FE analysis. At this stage,several design modifications are considered. The initiallyadopted rotor design is shown in Fig. 6(a), and the consideredalternative design adaptations are presented in Fig. 6(b)–(d).

Table I presents four different design cases to illustratethe influences of the structural adaptation on the rotor stressdistribution. Geometry case A represents the initial design[Fig. 6(a)] considered, and geometry cases B1–B3 are mod-ified geometry cases of Fig. 6(b). Associated stress plots forcases A and B3 are shown in Fig. 7(a) and (b), respectively.

As can be seen from the results, although the geometryin case A can provide the required magnetic performance,it experiences a high stress concentration localized at theedges [see Fig. 7(a)]. Thus, a round corner is introducedto minimize the stress as explained in Table I. It is worthnoting that the maximum stress is significantly reduced when15-mm-radius round corner edges are adopted as shown inFig. 7(b). Although the additional radius on the edges is ben-eficial for stress-level reduction, the arrangement reduces the


TABLE II

INFLUENCE OF THE BRIDGE THICKNESS AND MAGNET CROSS-SECTIONALAREA ON ROTOR STRESS

active magnet area; consequently, the magnetic performanceis also altered. Given that the stress limit is lower than theyield strength of the material, either the bridge can be furtherreduced or the magnet area can be increased by increasing itsdepth to enhance the air-gap flux density.

Table II shows a series of results obtained for various combi-nations of bridge thickness and magnet area. As expected, themaximum stress is more sensitive to the bridge thickness thanthe magnet cross-sectional area. In comparison with differentcases considered in the study, case C7, while it experiences amaximum stress of 1100 MPa with a safety factor of 1.32,fulfils the minimum safety factor set beforehand. Also, itis worth noting that among the arrangements (C1–C7), onlycase C7 satisfies the starting torque requirements. Therefore,case C7 is investigated further, adopting the arrangementsshown in Fig. 6(c) and (d). The stress distributions obtainedfor these cases are presented in Fig. 8. It is worth noting thatby introducing an air slot (case C71), or removing materialat the rotor outer periphery (case C72), the required torque atstarting is achieved while stress distribution is not significantlyaltered. This can be seen in Fig. 8. Finally, case C72 is selectedconsidering the slight reduction of maximum rotor stress.

B. Eddy Current Loss Analysis

Although the design is optimized to fulfill both magneticand structural stress constraints, it is important to ensure thatthe losses generated in the rotor satisfy the thermal limit.The losses are therefore estimated for different arrangementsof the following parameters: a different number of slits onthe rotor body, the space between the slits, and the slit depth.For the computation, 3-D FE electromagnetic analysis is used.The resultant eddy current losses for different arrangementsare gathered and analyzed. The obtained results are presentedin Table III and Fig. 9.

From Table III, it is also clear that the space betweenthe rotor slits has no influence on the induced eddy currentlosses. Conversely, the slit depth has significant influence, withlosses reducing with increasing slit depth. This relationship ismore apparent at shallower depths of slit; for example, theloss difference between the slit depths of 20 and 26 mm for

Fig. 8. Stress distribution in the IPM rotor. (a) Case C7. (b) Modified designwith air slot (case C71). (c) Modified design with material reduction on therotor surface (case C72). (Units are in megapascals.)

TABLE III

EDDY CURRENT LOSSES FOR DIFFERENT SLIT ARRANGEMENTS

the considered rotor is less than 3.5%, while the differencebetween 15 and 20 mm is 31.5%. Therefore, implementinga minimal slit thickness allows improving the rotor stiffnessand, thus, operational speed.

From Fig. 9, it can be seen that, as expected, the eddycurrent losses reduce with an increasing number of rotor silts.


Fig. 9. Rotor iron eddy current losses versus the number of slits on the rotor.

This is due to the reduction in the eddy current path thatbecomes smaller with an increasing number of slits. Selecting57 slits–which provides the rotor with a slit thickness of1 mm–results in a ∼17% (starting) torque reduction (for agiven current loading), compared with a completely solid rotor.A rotor with 19 slits (4-mm slice thickness) has a torquereduction of ∼3%. However, torque can be improved if theslit space is kept minimal. It is worth noting that the keydesign limitation comes from manufacturing practicality, so aminimal space between slits of 0.2 mm is selected based oncurrent manufacturing feasibility.

C. Thermal Analysis

In order to ensure that the design is working within theoperational temperature range, a thermal analysis is carriedout for different slit thicknesses. In the design, the spacebetween the slit and the slit depth are set to be 0.2 and 20 mm,respectively. Due to the high power density requirements in thedesign, the machine is loaded with a high current density. Oilcooling is therefore adopted for the stator, in which the statorregion is isolated from the rotor region and flooded with oil,allowing the rotor to run dry and maintain minimal windagelosses. A lumped parameter thermal model is used to predictthe thermal performance. For accuracy, a 1/6 cross sectionof the rotor is modeled initially using Computational FluidDynamics software with the predicted heat transfer coefficientlater implemented in the lumped model. The detailed thermalanalysis adopted for the design can be found in [24]. It is worthhighlighting that the temperature should be limited within therotor as the cooling is available only at the stator.

The obtained temperature distributions for 3-mm slittedrotor are shown in Fig. 10. A comparison of the results at slitthicknesses of 2-, 3-, and 4-mm slitted rotors is presented inTable IV. From the results, it is obvious that the temperature–and therefore losses–varies as a function of slit thickness.It is also worth noting that the temperature of the magnet(Recoma grade 30S, samarium cobalt [25]) is lower than themagnet’s maximum thermal limit (350 °C). This confirms thatthe design is thermally viable.

D. Rotor Dynamic Analysis

Since the rotor is slitted into several disks, the dynamicprofile of the rotor is expected to change. Rotor dynamic

Fig. 10. Temperature distributions inside the 3-mm slitted rotor machine.

TABLE IV

TEMPERATURE COMPARISON BETWEEN DIFFERENT SLIT THICKNESSES

TABLE V

EFFECTS OF DEPTH OF SLITTING AND SPACE BETWEEN THE STACKS

ON DYNAMIC PERFORMANCES OF ROTOR

analysis is therefore performed for both 3- and 4-mm slittedrotors, with different cases of slit depth, and compared withan unmodified whole rotor. The results obtained are listed inTable V.


Fig. 11. Dynamic response of (a) solid rotor, (b) 3-mm slitted rotor, and(c) 4-mm slitted rotor (S4). (Units are in millimeters.)

From Table V, it is evident that while the mass of the magnetinfluences the natural frequencies, they are well above theoperating speed of the machine. Slicing the rotor decreasesits stiffness and consequently reduces the natural frequenciesas illustrated in cases S2–S4. This effect is not large enough tochange the mode shape of shaft as shown in Fig. 11. It is clearfrom Fig. 11 that the rotor shaft does affect the stiffness andreduces the natural frequencies with increasing slit depth. Alsoit is evident here that there are no significant effects on rotorstiffness due to changes in space distance between the slits.

Although the natural frequency of the design is higher thanthe operating speed, it is important to ensure that the bearingstiffness is sufficient. The design that has a rotor shaft with26-mm slit depth and 0.4-mm space between the slitsis therefore investigated by setting a bearing stiffness of1E8 N/m at both shaft ends. Fig. 12 shows the obtainedCampbell diagram for the considered bearing-rotor-shaft case.From the results, it can be seen that with a bearing stiffness of1E8 N/m at both shaft ends, the critical speed is well belowthe natural frequency, and thus, the rotor is stable.

In conclusion, the proposed solid rotor topology is feasiblemagnetically and structurally. The design is also thermally fea-sible if the rotor losses are minimal as expected. In Section VI,an experiment is carried out to verify the eddy current lossesassociated with the solid materials being considered for thisapplication.

VI. EXPERIMENTAL CHARACTERIZATION

OF SOLID ROTOR

The key scope of this section is to evaluate the eddy currentlosses associated with the proposed slitted rotor in comparison

Fig. 12. Campbell diagram for the rotor with 0.4-mm space between thestack and 26-mm milling depth under a bearing stiffness of 1E8 N/m (mode–bending mode; BW–backward; and FW–forward).

Fig. 13. Two different rotor arrangements. (a) Slitted body. (b) Completesolid body. (c) Prepared samples for testing.

with a solid rotor body. This allows confirming that theproposed slitted concept minimizes the losses, and thus, theconcept is thermally viable. Two different solid arrangementsthat have the same dimensions (outer diameter, inner diameter,and stack length) are adopted for the measurement. These area complete solid body and a slitted body as shown in Fig. 13.A representative stainless steel (grade 431) sample availablein the laboratory is adopted for the test. The slitted body hasfive slits, a width of 2 mm, and a depth of 20 mm from theouter surface of the rotor.

For the loss measurement, a state-of-the-art [26] testingfacility that is available in-house was used. The setup specifi-cations are shown in Table VI. It is worth noting that the oper-ating temperatures depend on the furnace capability and not onthe capability of the measurement facility. The actual pressure(MPa) applied to the samples is subject to the dimensions ofthose samples. The samples prepared for the measurement areshown in Fig. 13(c) where the test samples are wound usingtwo windings, the primary and the secondary, to magnetizethe steel and measure the dropped and induced voltages acrossthe medium, respectively. Based on the gathered voltage and


TABLE VI

SPECIFICATIONS AND CAPABILITIES OF THE MEASUREMENT FACILITY

Fig. 14. Loss comparison of different rotors for 40- and 50-Hz sinusoidalexcitation.

TABLE VII

LOSSES’ COMPARISON BETWEEN THE TWO DIFFERENTROTOR ARRANGEMENTS

current data, the losses are estimated. The obtained lossesare presented in Fig. 14. The measurements are limited toa frequency of 50 Hz, since the current (and consequentlyvoltage) providing the flux density level in the sample isalso limited. Testing the samples at a higher frequency whilemaintaining the flux density level requires a larger supply thanthat is available. However, the losses obtained at low frequencyprovide an indication of the loss reduction due to slitting.

Fig. 14 shows the ac loss comparison between the solidbody and slitted body for excitation frequencies of 40 and50 Hz. As expected, there is a considerable reduction in thelosses associated with the slitted body–an average reductionof 48%–compared with the complete solid body.

Table VII shows the estimated losses under load for theconsidered IPM machine with two different rotor arrangements(solid rotor and rotor with five slits) based on the measuredmaterial (stainless steel 431) loss data. From the results, it

is evident that the losses associated with the slitted rotor aresignificantly lower than the complete solid rotor. This furtherconfirms that the slitted rotor topology effectively minimizesthe eddy current losses.

VII. CONCLUSION

In this paper, a novel solid rotor topology that can cope withhigh structural stress has been proposed for IPM machines.The stress distributions and dynamic performances associatedwith such rotor concepts are analyzed using FE. An optimiza-tion study on IPM shape and the curvature angle has beendone to improve the structural performance. To minimize theeddy current losses, a partly slitted solid rotor that shorts theeddy current paths has been investigated. It was shown thatthe depth of slit has to be limited because of the reductionin both stiffness and natural frequency with an increase inslit depth. It was also shown that space between the rotorslices does not have any influence on rotor stiffness, and islimited only by the manufacturability of such a design. Eddycurrent analysis confirmed that the losses are a function ofthe thickness of the rotor slices and a tradeoff between theperformance and the losses is required. It was also shownthat the design is thermally viable and that it is important toreduce the rotor losses to a minimum in order to maintainthe temperature within limits, by taking into account availablecooling. The results obtained from the two tested stainless steelsamples–solid rotor construction and a slitted arrangement–further confirmed that the proposed slitted approach reducesthe eddy current losses significantly, and thus, the topologycan effectively be implemented within the IPM machine.

REFERENCES

[1] B. Mecrow, J. Cullen, and P. Mellor, “Editorial—Electrical machines anddrives for the more electric aircraft,” IET Electr. Power Appl., vol. 5,no. 1, pp. 1–2, Jan. 2011.

[2] A. Boglietti, A. Cavagnino, A. Tenconi, and S. Vaschetto, “The safetycritical electric machines and drives in the more electric aircraft:A survey,” in Proc. 35th Annu. Conf. IEEE Ind. Electron. (IECON),Nov. 2009, pp. 2587–2594.

[3] C. Liu, K. T. Chau, and X. Zhang, “An efficient wind–photovoltaichybrid generation system using doubly excited permanent-magnet brush-less machine,” IEEE Trans. Ind. Electron., vol. 57, no. 3, pp. 831–839,Mar. 2010.

[4] A. Di Gerlando, G. Foglia, M. F. Iacchetti, and R. Perini, “Axial fluxPM machines with concentrated armature windings: Design analysis andtest validation of wind energy generators,” IEEE Trans. Ind. Electron.,vol. 58, no. 9, pp. 3795–3805, Sep. 2011.

[5] M. Pinilla, “Performance improvement in a renewable energy directdrive permanent magnet machine by means of soft magnetic compositeinterpoles,” IEEE Trans. Energy Convers., vol. 27, no. 2, pp. 440–448,Jun. 2012.

[6] P. Arumugam, T. Hamiti, C. Brunson, and C. Gerada, “Analysis of verti-cal strip wound fault-tolerant permanent magnet synchronous machines,”IEEE Trans. Ind. Electron., vol. 61, no. 3, pp. 1158–1168, Mar. 2014.

[7] X. Jiang, W. Huang, R. Cao, Z. Hao, and W. Jiang, “Electric drivesystem of dual-winding fault-tolerant permanent-magnet motor foraerospace applications,” IEEE Trans. Ind. Electron., vol. 62, no. 12,pp. 7322–7330, Dec. 2015.

[8] B. K. Bose and P. M. Szczesny, “A microcomputer-based control andsimulation of an advanced IPM synchronous machine drive system forelectric vehicle propulsion,” IEEE Trans. Ind. Electron., vol. 35, no. 4,pp. 547–559, Nov. 1988.

[9] A. Choudhury, P. Pillay, and S. S. Williamson, “Modified DC-linkvoltage balancing algorithm for a 3-level neutral point clamped (NPC)traction inverter based electric vehicle PMSM drive,” in Proc. 39th Annu.Conf. IEEE Ind. Electron. Soc. (IECON), Vienna, Austria, Nov. 2013,pp. 4660–4665.


[10] S.-M. Sue and C.-T. Pan, “Voltage-constraint-tracking-based field-weakening control of IPM synchronous motor drives,” IEEE Trans. Ind.Electron., vol. 55, no. 1, pp. 340–347, Jan. 2008.

[11] S. Y. Jung, C. C. Mi, and K. Nam, “Torque control of IPMSM inthe field-weakening region with improved DC-link voltage utilization,”IEEE Trans. Ind. Electron., vol. 62, no. 6, pp. 3380–3387, Jun. 2015.

[12] T. Sun and J. Wang, “Extension of virtual-signal-injection-based MTPAcontrol for interior permanent-magnet synchronous machine drives intothe field-weakening region,” IEEE Trans. Ind. Electron., vol. 62, no. 11,pp. 6809–6817, Nov. 2015.

[13] E. Al-Nabi, B. Wu, N. R. Zargari, and V. Sood, “Sensorless controlof CSC-fed IPM machine for zero- and low-speed operations usingpulsating HFI method,” IEEE Trans. Ind. Electron., vol. 60, no. 5,pp. 1711–1723, May 2013.

[14] D. Dorrell, L. Parsa, and I. Boldea, “Automotive electric motors,generators, and actuator drive systems with reduced or no permanentmagnets and innovative design concepts,” IEEE Trans. Ind. Electron.,vol. 61, no. 10, pp. 5693–5695, Oct. 2014.

[15] A. M. El-Refaie, “Fractional-slot concentrated-windings synchronouspermanent magnet machines: Opportunities and challenges,” IEEETrans. Ind. Electron., vol. 57, no. 1, pp. 107–121, Jan. 2010.

[16] R. Ni et al., “Efficiency enhancement of general AC drive system byremanufacturing induction motor with interior permanent-magnet rotor,”IEEE Trans. Ind. Electron., vol. 63, no. 2, pp. 808–820, Feb. 2016.

[17] M. N. Uddin and M. A. Rahman, “High-speed control of IPMSM drivesusing improved fuzzy logic algorithms,” IEEE Trans. Ind. Electron.,vol. 54, no. 1, pp. 190–199, Feb. 2007.

[18] P. Arumugam et al., “Comparative design analysis of permanent magnetrotor topologies for an aircraft starter-generator,” in Proc. IEEE Int.Conf. Intell. Energy Power Syst. (IEPS), Jun. 2014, pp. 273–278.

[19] M. Barcaro, G. Meneghetti, and N. Bianchi, “Structural analysis of theinterior PM rotor considering both static and fatigue loading,” IEEETrans. Ind. Appl., vol. 50, no. 1, pp. 253–260, Jan./Feb. 2014.

[20] K. Baoquan, L. Chunyan, and C. Shukang, “Flux-weakening-characteristic analysis of a new permanent-magnet synchronous motorused for electric vehicles,” IEEE Trans. Plasma Sci., vol. 39, no. 1,pp. 511–515, Jan. 2011.

[21] C. Zuhou, “Variable excitation permanent magnet synchronous motor,”Google Patent 2011 10 109 156, Aug. 31, 2011. [Online]. Available:http://www.google.com.gt/patents/CN102170211A?cl=en

[22] A. Tessarolo, M. Mezzarobba, and R. Menis, “A novel interior per-manent magnet motor design with a self-activated flux-weakeningdevice for automotive applications,” in Proc. 20th Int. Conf. Elect.Mach. (ICEM), Sep. 2012, pp. 2603–2609.

[23] M. S. K. Hochhalter, “Permanent magnet motor or actuator with fieldweakening capability,” U.S. Patent 2010 02 137 79, Aug. 26, 2010.

[24] Z. Xu, A. L. Rocca, S. J. Pickering, C. Eastwick, C. Gerada,and S. Bozhko, “Mechanical and thermal design of an aero-engine starter/generator,” in Proc. IEEE Int. Electr. Mach. DrivesConf. (IEMDC), May 2015, pp. 1607–1613.

[25] ARNOLD Magnetic Technologies. RECOMA Samar-ium Cobalt Magnets, accessed on Jun. 2015. [Online].Available: http://www.arnoldmagnetics.com/Portals/0/Files/RECOMA/Recoma%20Combined%20-%20160301.pdf?ver=2016-03-01-134105-057

[26] W. L. Soong, BH Curve and Iron Loss Measurements for MagneticMaterials (Power Engineering Briefing Note Series). Sch. Elect. ElectronEng., Univ. Adelaide, Australia, May 12, 2008.

Puvan Arumugam (M’11) received theB.Eng. degree (Hons.) in electrical and electronicsengineering in 2009, and the Ph.D. degree inelectrical machines and drives from The Universityof Nottingham, Nottingham, U.K., in 2013.

He was a Researcher with the Power Electronics,Machines, and Control Group, The University ofNottingham, where he was involved in electricaircraft propulsion. He is currently a SeniorProject Engineer with the Force Engineering Ltd.,Shepshed, U.K. His current research interests

include electrical machines and drives, electromechanical devices andsystems, and analytical computation of electromagnetic fields.

Dr. Arumugam was a recipient of the Hermes Fellowship supported byTechnology Transfer Office, The University of Nottingham in 2014.

Zeyuan Xu received the Ph.D. degree in mechani-cal engineering from the University of Manchester,Manchester, U.K., in 2002.

He was a Researcher with Brunel University,Uxbridge, U.K., and The University of Nottingham,Nottingham, U.K. His current research interestsinclude turbulent thermos-fluid flow, heat transferenhancement, thermal management of advance elec-trical motor and power electronics, and mechanicaldesign of high-speed electrical machine.

Antonino La Rocca received the B.Eng. andM.Eng. degrees in mechanical engineering from theUniversity of Palermo, Palermo, Italy, in 2010 and2011, respectively, and the Ph.D. degree in mechani-cal engineering from The University of Nottingham,Nottingham, U.K., in 2016.

He is currently a Researcher with the Fluids andThermal Engineering Research Group, The Univer-sity of Nottingham. His current research interestsinclude fluid-dynamics, heat transfer, and analyti-cal and numerical thermal modeling of electrical

machines.

Gaurang Vakil received the B.E. degree in electri-cal engineering from Saurashtra University, Rajkot,India, in 2006, and the M.Tech. degree from NirmaUniversity, Ahmedabad, India, in 2008. He is cur-rently pursuing the Ph.D. degree with the PowerElectronics, Machines and Drives Group, IIT Delhi,New Delhi, India, with a focus on variable speedgenerator design for renewable energy applications.

He is a Research Associate with the Power Elec-tronics, Machines and Controls Group, The Univer-sity of Nottingham, Nottingham, U.K. His current

research interests include analytical modeling and design optimization ofelectrical machines, optimizing electric drive-train for pure electric and hybridvehicles, high-power density machines, magnetic material characterization,and high-performance electrical machines for transport, traction, and renew-able energy applications.

Mr. Vakil was a recipient of the Gold Medal for the M.Tech. degree fromNirma University.

Matthew Dickinson received the B.Eng. degree(Hons.) in mechanical engineering from De MontfortUniversity, Leicester, U.K., in 2008.

He is currently a Project Engineer with ForceEngineering Ltd., Shepshed, U.K., where he isinvolved in the thermal and mechanical design oflinear motors. His current research interests includethermal management, analytical computational toolsfor linear motors, and advanced materials for elec-trical machines.

Emmanuel Amankwah received the B.Sc. degreein electrical and electronics engineering from theKwame Nkrumah University of Science and Tech-nology, Kumasi, Ghana, in 2006, the M.Sc. degree inelectrical engineering and the Ph.D. degree in electri-cal and electronics engineering from The Universityof Nottingham, Nottingham, U.K., in 2009 and 2013,respectively.

He was with the Electricity Company of Ghana,Ghana, as a Design Engineer, from 2006 to 2008.Since 2013, he has been a Research Fellow in

emerging technologies for HVDC power transmission with the Faculty ofEngineering, Power Electronic Machines and Control Research Group, TheUniversity of Nottingham. His current research interests include power elec-tronics for grid integration and motor drive control.


Tahar Hamiti was born in Larbaâ Nath Irathen,Algeria, in 1979. He received the Ingénieurd’Etat degree in automatic control systems fromthe Mouloud Mammeri University of Tizi-Ouzou,Tizi-Ouzou, Algeria, and the Ph.D. degree in elec-trical engineering from the University of Nancy I,Nancy, France.

He was a Research Fellow and a Lecturer withthe Power Electronics, Machines and Control Group,The University of Nottingham, Nottingham, U.K.,from 2010 to 2015. In 2015, he joined VEDECOM,

a French institute for energy transition, and was involved in novel electricalmachines for electric and hybrid vehicles. His current research interestsinclude modeling, optimal design, and control of high-performance electricalmachines for transportation applications and power generation.

Serhiy Bozhko (M’96) received the M.Sc. andPh.D. degrees in electromechanical systems fromKyiv Polytechnic Institute, Kiev, Ukraine, in 1987and 1994, respectively.

He has been with the Power Electronics, Machinesand Controls Research Group, The University ofNottingham, Nottingham, U.K., since 2000. He hasbeen a Principal Research Fellow and is currently anAssociate Professor of Aircraft Electrical Systems.He has been involved in leading several EU- andindustry-funded projects in the area of aircraft elec-

tric power systems, including control and stability issues, power management,and advanced modeling and simulations methods.

Chris Gerada (M’05) received the Ph.D. degreein numerical modeling of electrical machines fromThe University of Nottingham, Nottingham, U.K.,in 2005.

He was a Researcher with The Universityof Nottingham, where he was involved inhigh-performance electrical drives and thedesign and modeling of electromagneticactuators for aerospace applications. Since2006, he has been the Project Manager of GEAviation Strategic Partnership, Nottingham.

In 2008, he was appointed as a Lecturer in Electrical Machines, an AssociateProfessor in 2011, and a Professor with The University of Nottingham, in2013. He serves as an Associate Editor of the IEEE TRANSACTIONS ONINDUSTRY APPLICATIONS and the Chair of the IEEE Industrial ElectronicsSociety Electrical Machines Committee. His current research interests includethe design and modeling of high-performance electric drives and machines.

Stephen J. Pickering received the B.Sc. andPh.D. degrees in mechanical engineering fromThe University of Nottingham, Nottingham, U.K.,in 1979 and 1984, respectively.

He joined The University of Nottingham as aLecturer in 1988, where he is currently a Hives Pro-fessor of Mechanical Engineering with the Facultyof Engineering. He has involved in thermos-fluidsand his current research interests include the coolingof electric machines for over twenty years.

ieee transactions on transportation …...ieee transactions on transportation electrification, vol....

Documents