Politecnico di Torino

Porto Institutional Repository

[Proceeding] Design of a line-start synchronous reluctance motor

Original Citation:M. Gamba;G. Pellegrino;A. Vagati;F. Villata (2013). Design of a line-start synchronous reluctancemotor. In: 2013 International Electric Machines & Drives Conference, Chicago, USA, Maggio 2013.pp. 648-655

Availability:This version is available at : http://porto.polito.it/2518606/ since: January 2016

Publisher:IEEE

Published version:DOI:10.1109/IEMDC.2013.6556163

Terms of use:This article is made available under terms and conditions applicable to Open Access Policy Article("Public - All rights reserved") , as described at http://porto.polito.it/terms_and_conditions.html

Porto, the institutional repository of the Politecnico di Torino, is provided by the University Libraryand the IT-Services. The aim is to enable open access to all the world. Please share with us howthis access benefits you. Your story matters.

Publisher copyright claim:c© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtainedfor all other uses, in any current or future media, including reprinting/republishing this material foradvertising or promotional purposes, creating new collective works, for resale or redistribution toservers or lists, or reuse of any copyrighted component of this work in other works

(Article begins on next page)

Abstract—The design of line-start Synchromachines is proposed, based on the state of tdesigned for closed loop, vector control. Thovercoming the dilemma between synchronizatsteady state performance are given, with exambad design choices. The proposed solutions coelement analysis with a competitor induction mefficiency and power factor. The finite elemenfinally validated via experiments on a laborator

Index Terms – Synchronous Reluctance MMotor, Line Start Motor, High Efficiency MoDrives.

I. INTRODUCTION Line-start synchronous motors are adopted

line-supplied Induction Motors (IM) due efficiency [1-3]. All the cited examples refmagnet (PM) line start motors, and this has bintroduction of rare-earth PMs back in volatility of rare earth materials prices over tyears has led to reconsider the Synchronous Rmachine as a viable alternative to PM lineconstant speed applications where line sapplicable. Line-Start SyR (LSSyR) machinand adopted in the ‘60s and the ‘70s [4-5]. Amachines were voltage supplied by mefrequency inverters to obtain exact speedspossible at the time with the induction motorLately, the diffusion of vector controlled Iprecise speed control, on the one hand, and ostart machines with a higher torque density, ohave led to abandon LSSyR motors.

In retrospect, being the literature about LSall the improvements of SyR motor design enow and the ‘70s have never been tesapplications. The research upon SyR mocontrolled drives has produced up to date mu[7], having better saliencies than the ones adin the literature, and disclosing a potential foand a better power factor at synchronous spee

This work proposes the design of a LSSyRapplication, based on a SyR rotor with four fwith aluminum completely and short-circuitethe stack to from the rotor cage. The deslaminations is based on the state of the art odesign, with some little modifications that s

All authors are with the Politecnico di Torino, DeTorino, Italy (e-mail: gianmario.pellegrino@polito.it)

M. Gamba, G. Pellegrino, S

Design of a Line-S

onous Reluctance the art of motors he guidelines for

tion capability and mples of good and ompared via finite motor in terms of

nt calculations are ry prototype.

Machine, Induction otor, Fixed Speed

d in alternative to to their better

fer to permanent een true since the the 1970s. The

the last couple of Reluctance (SyR)

e-start motors for start motors are nes were studied At that time, such eans of variable s, that were not r counterparts [6]. IM drives with a of PM based line

on the other hand,

SSyR rather aged, emerged between sted in line-start otors for vector-

multi-barrier rotors dopted for LSSyR or a higher torque ed.

R motor for pumps flux barriers filled ed at both ends of sign of the rotor of SyR machines show to improve

epartment of Energy,

the starting capability of the machifor obtaining the best compromisepull-out torque values are given. Ththe synchronization capability intenthat can be put into step. The pullperformance at synchronous speedload applicable at synchronismcapabilities are in contrast with eachand a compromise is given.

A lumped parameters model of thsimulated in the time domain ttransients. From the same model expressed analytically, as a functionthe voltage vector, defining the syrotor. The lumped parameters apprbut it is not very accurate in termstorque and the inertia that can be acElement Analysis (FEA) simulatiomotion type are then used, and the different motors under test is evalutotal inertia to be synchronized. Aare compared at steady state with ththe same frame and the same stator Last, a prototype SyR machine, eqwith welded copper bars partiallytested for checking the validity of FEA.

Figure 1. Line Start SyR motors rotorssynduction motor-like solution [4]. LS2filled with aluminum (dark grey areas).

solution.

Senior Member, IEEE, A. Vagati, Fellow, IEEE, and

tart Synchronous Reluct

ine. The design guidelines e between the pull-in and he pull-in torque represents nded as the maximum load l-out torque represents the d, meaning the maximum

m. Pull-in and pull-out h other, to a certain extent,

he motor is presented, and o investigate the pull-in the steady state torque is

n of the slip speed between ynchronous speed, and the roach gives precious hints s of evaluating the pull-in ctually put into step. Finite ons of the transient with pull into step curve of the uated as a function of the

All tested LSSyR solutions he an IM competitor having

laminations and windings. quipped with a cage made y filling the saliencies, is

the transient with motion

(LS1)

(LS2)

(LS3)

s under investigation: LS1 is a 2 is a state of the art SyR rotor . LS3 is the proposed LSSyR

F. Villata

tance motor

II. MODELING OF THE LSSYR MA

The dynamic model of the LSSyR machineconcurrent presence of rotor saliency, as in and a short circuited rotor cage, as in an InduThe dq reference frame, synchronous to the roFig. 2. The rotor speed, in electrical radians, ivector, also in Fig. 2, is imposed by the AC the angular frequency ω and the synchronwhere p is the number of pole pairs. In thevoltage slips at ω − ωr and the slip s is defined

ωω−ω

= rs (1)

The phase angle δ of the voltage vector iFig. 2.

β q

Figure 2. Definition of the stator and rotor synchframes. Definition of the synchronous s

The rotor cage is not housed in usual rosame and regularly displaced along the airgawould be in an asynchronous motor. On the cconductors here are aluminum bars that fill the SyR rotor completely, as in the examples

A. Dynamic model With reference to the circuital model repo

the d and q axes, the electrical equations of th

srs

sss jdt

diRv λω+

λ+=

dtd

iR

R rr

rq

rd λ+⋅⎥

⎦

⎤⎢⎣

⎡=

00

0

dtλdiR m

fefe +=0 (

The subscript s stands for “stator” variasubscript r stands for “rotor”. The non isotrocage bars is reflected into the rotor parameteand then the rotor resistances are different forin (3). The magnetic model is:

mmq

mdsσss i

LL

iLλ ⋅⎥⎦

⎤⎢⎣

⎡+=

00

ACHINE e accounts for the

a SyR machine, uction Motor [1]. otor, is defined in is ωr. The voltage mains that define nous speed ω/p,

e rotor frame, the d according to:

is also defined in

d

w r t

α

hronous reference speed ω.

otor slots, all the ap periphery, as it contrary, the cage the saliencies of

of Fig. 1.

orted in Fig 3 for he motor are:

(2)

(3)

(4)

ables, whilst the opic shape of the ers of the model, r the d and q axes

(5)

rσrq

σrdr

Li

LL

λ ⎢⎣

⎡+⋅⎥

⎦

⎤⎢⎣

⎡=

00

rsm iii +=

Where Lmd and Lmq are the magneLσr are the stator and rotor leakageAs for the rotor resistances, also theare not equal in d and q in (6). Basithe circuit related to the rotor reflect

Figure 3. Dynamic equivalent circuicomponents

Moreover, the magnetizing induboth the magnetizing current compand cross-saturation, as usual for linear relationship is expressed magnetizing flux components tocomponents, as reported in Fig. 4. Tmapping the machine with static FE

Figure 4. Non linear relationship betwcurrent components, in the roto

The electromagnetic torque andare:

( ssem ipT ∧λ⋅= 23

loaemtot TTdtdJ −=ω

Where Jtot accounts for the motoris the load torque.

B. Steady state torque as a functionAs reported in [1] the avera

components during run up can be de

mmq

md iL

L⋅⎥

⎦

⎤0

0 (6)

(7)

etizing inductances. Lσs and inductances, respectively.

e rotor leakage inductances cally, all the parameters of t the rotor anisotropy.

it of LSSyR machine, in dq s.

uctances are a function of ponents, due to saturation SyR machines. The non in tables, relating the

o the respective current The tables are obtained by A.

ween the magnetizing flux and or reference frame dq.

d the mechanical equation

) (8)

d (9)

r and load inertia and Tload

n of the slip speed ge and pulsating torque escribed via a quasi steady

state analysis. The first assumption is that the electric and magnetic phenomena are faster than the mechanical transients, so that the slip speed can be assumed constant time by time, while the voltage vector steadily rotates around the rotor at slips speed. The steady state electrical condition corresponds to replace the time derivatives d/dt in (2)-(4) with the operator j(sω), valid for steady state sinusoidal variables, as they were all phasors. The angular frequency of the phasors is sω because of the chosen rotor frame. The electrical equations (2)-(3) become (10) and (11) respectively:

⎪⎩

⎪⎨⎧

Λω+Λω+=

Λω−Λω+=

sdrsqqssq

sqrsddssd

jsIRV

jsIRV (10)

⎩⎨⎧

Λω+=Λω+=

rqrqr

rdrdr

jsIRjsIR

00

(11)

Where capital letters indicate phasors. The torque expression (8) requires the stator current and flux linkage components to be expressed as dq phasors and then the vector cross-product to be calculated. To eliminate the rotor current from equations, this can be expressed as a function of the stator current by manipulation of (10) with (6) and (7).

⎪⎪⎩

⎪⎪⎨

⎧

ω+⋅ω

−=

ω+⋅ω

−=

sqrqrq

mqrq

sdrdrd

mdrd

ILjsR

LjsI

ILjsR

LjsI

(12)

Where Lrd = Lσr + Lmd, Lrq = Lσr + Lmq. The rotor current (12) is eliminated from equation (5) and the stator flux to current phasor relationship is found.

⎪⎪⎪

⎩

⎪⎪⎪

⎨

⎧

⋅=⋅⎟⎟

⎠

⎞

⎜⎜

⎝

⎛

ω+ω

−=Λ

⋅=⋅⎟⎟⎠

⎞⎜⎜⎝

⎛

ω+ω

−=Λ

sqpqsqrqrq

mqsqsq

sdpdsdrdrd

mdsdsd

IZILjsR

LjsL

IZILjsR

LjsL

2

2

(13)

As for the rotor, also for the stator Lsd = Lσs + Lmd, Lsq = Lσs + Lmq. The two impedances Zpd and Zpq in (13) are called operational impedances: they are complex numbers intended as phasor operators, and their values are a function of the slip frequency sω or, in other words, of the actual rotor speed. Yet, the stator current phasors have to be calculated. The voltage to current relationship is obtained by manipulation of (10) and (13).

( )( ) ⎥

⎥⎦

⎤

⎢⎢⎣

⎡⋅

⎥⎥⎦

⎤

⎢⎢⎣

⎡

ω+ω−ω−−ω+=

⎥⎥⎦

⎤

⎢⎢⎣

⎡

sq

sd

pqspd

pqpds

sq

sd

II

ZjsRZ)s(Z)s(ZjsR

VV

11 (14)

The voltage dq phasors are:

⎪⎩

⎪⎨⎧

δ+ω⋅=

δ+ω⋅−=

)tscos(V̂V

)tssin(V̂V

sq

sd

0

0 (15)

Being the voltage phase angle δ referenced to the q axis of the rotor, as reported in Fig. 2. When the slip speed is not zero, the d and q components of the voltage vector are then phasors in time quadrature, regardless of the term δ0, that is the load angle at synchronous speed:

⎪⎩

⎪⎨⎧

=

=

V̂V

V̂jV

sq

sd (16)

The inverse of (14), finally, expresses the stator current phasors:

( )( ) ⎥

⎥⎦

⎤

⎢⎢⎣

⎡⋅

⎥⎥⎦

⎤

⎢⎢⎣

⎡

ω+ω−−ω−ω+⋅=⎥

⎦

⎤⎢⎣

⎡

V̂V̂j

ZjsRZ)s(Z)s(ZjsR

DII

pdspd

pqpqs

csq

sd

111 (17)

where

( ) ( ) pqpdpqpdssc ZZsZZRjsRD 22 21 ω−++ω+= (18)

By substitution of (17) into (13), also the stator flux components are determined. The amplitude and phase of all the phasors is a function of the slip angular frequency sω. In Cartesian form, they are then:

⎪⎪

⎩

⎪⎪

⎨

⎧

+=Λ+=Λ

+=+=

jhgjfe

jdcIjbaI

sq

ds

sq

sd

(19)

Recalling once more the torque expression (8), and substituting (13), (17) and (19), after a cumbersome manipulation, the quasi steady state torque is:

( )α−⋅ω⋅⋅+= tscosTTT relcagee 2 (20)

where

⎥⎦⎤

⎢⎣⎡ −−+⋅=

223 bhgadfcepTcage (21)

( ) ( )2243 ahgbcfedgadfbhcepTrel −−++−−+⋅= (22)

( )( ) ⎟⎟⎠

⎞⎜⎜⎝

⎛−−+−−+=α

gadfbhcecfedahgbarctan (23)

Tcage comes from the rotor cage currents, it is constant at a given slip speed, and its value depends on the rotor slip, as in a induction motor. Trel is the amplitude of the torque component produced by the rotor saliency, that is alternative and pulsating

at two times the slip angular frequency. α is the phase angle of the pulsating torque component.

C. Discussion of the two torque components When dealing with the start-up transient of the LSSyR

motor, all of a sudden switched to the AC mains, the asynchronous cage torque acts as pull-up torque, meaning that it is the one accelerating the motor from standstill towards synchronism. The quasi steady-state characteristic of this torque component, as a function of the rotor speed, resembles the torque characteristic of a voltage supplied IM, as represented in Fig. 5.

The reluctance torque is an alternating component at all speeds but synchronism, pulsating at twice the slip frequency during the starting phase of the motor. The steady-state envelope of Tcage-Trel to Tcage+Trel is represented in Fig. 5. Also the peak value Trel is a function of the slip speed, as the height of the band around Tcage is not the same at all speeds. In particular, Tcage+Trel at synchronous speed coincides with the pull-out torque of the machine: as usual for synchronous machines this is the maximum load torque that is applicable without loss of synchronism. The pull-out torque is normally much higher than the rated torque of the motor. However, having a high pull-out torque over rated torque ratio is an important figure of merit because it indicates that the motor has a high power factor at synchronous speed.

Figure 5. Quasi steady state characteristic of Tcage and of the Trel

envelope for a LSSyR motor

D. Results from the transient and steady-state models When line-starting a speed-dependant load such as a pump,

the torque-speed transient is of the kind of the one in Fig. 6. after the initial electrical transient is extinguished, the instantaneous torque is fairly bounded within the steady state envelope. Fig. 7 shows the torque and speed transient of Fig. 6 as a function of time.

Figure 6. Start-up transient, with a load torque quadratic with speed.

(a)

(b)

Figure 7. Torque (a) and speed (b) transients. Same simulation as Fig. 6.

E. Pull-in torque characteristic The pull-in torque is the maximum load torque under

which the motor puts a given inertia into synchronism. It is then a measure of the synchronization capability of the motor and it varies inversely with the total system inertia, as in the characteristic reported in Fig. 8 for the LS2 motor of Fig. 1. The motor ratings are summarized in Table I.

For each value of inertia, the points in Fig. 8 are evaluated by repeating the transient simulations for different load levels, where the load on the Nm abscissa is intended as the load torque at synchronous speed, assuming that the load at non synchronous speed is proportional to the square of the actual rotor speed. Two curves in Fig. 8 refer to the lumped parameters model and one to the transient FEA analysis. The latter shows the best results, while the former ones show the imprecision of the lumped parameters model, and the sensitivity to the correct determination of the rotor parameters. One curve refers to the rotor parameters at zero speed, the other one close to synchronism, with the method in [8], and both cases are far from being accurate.

Figure 8. Pull-in torque characteristic as a function of the total inertia, for the motor LS2 of Fig. 1..

III. DESIGN CRITERIA AND EXAM

A. Maximization of the pull-in capability The aim of a LSSyR design is twofold:

efficiency and power factor at synchronism apull-in torque close to the rated torque overvalues inertias to be pulled-into stepcharacteristics of Fig. 8 have been obtained lumped parameters model presented ad imprecision demonstrated by the analytical mrelated to the fact that some key aspects ssaturation and skin effect in the rotor bars aalso that the determination of key parametersresistances and leakage inductances is not trierrors [8]. Yet, the model can be useful fordesign, as summarized by the following consi1. The performance at synchronous sp

power factor, pull-out torque) is dosaliency ratio, and saliency maximizatdirection to go.

2. Reducing the d and q rotor resistancpull-in torque.

The example of Fig. 9 shows that halvresistance of an example motor produces torque, meaning a steeper characteristic arospeed and a higher value of the maximum totime. This turns to be advantageous in capability, meaning that the inertia that can bwill be higher , once the rotor resistance has for the torque characteristic of an IM, the flipthe rotor resistance is the decrease of the stalis only partially a problem because the included in Fig. 9, tends to increase the rotor slip values, including stall, helping the stahigh. In other words, the cage torque atunderestimated in Fig. 9.

Figure 9. Effect of reducing the q axis rotor resteady-state torque characteristic of a LSSyR. Bl

R’rd=Rrd, R’rq=1.36Rrq.

B. Evolution from rotor LS1 to rotor LS3 From point (1) at subsection III.A, it tur

motors with up to date rotors, having a highthe ones at the time of [4], can have a bettersteady state, meaning a higher pull-out torqufactor and then a lower torque to absorbed then efficiency. From Fig. 9 it follows that a

MPLES

: to have a high and also to have a r a wide range of p. The pull-in according to the section II. The

model in Fig. 8 is such as magnetic are neglected, and s such as the rotor ivial and prone to r orienteering the iderations: peed (efficiency, ominated by the tion is the correct

ces improves the

ving the q rotor a better pull-up

ound synchronous orque at the same terms of pull-in

be pulled into step been reduced. As

p side of reducing ll torque, but this skin effect, not resistance at high

all torque to stay t stall is by far

esistance on the lack lines refer to

rns out that SyR her saliency than r performance, at

ue, a better power current ratio and

all available rotor

space must be filled with alumresistances as low as possible. This literature, where the primordial LSslotted cages with limited cross sefavor of solutions with larger fsaliencies filled with aluminum [6motor in [4].

The starting point of the design ethree rotors in Fig. 1 All three motaken from a 2.2 kW, four pole, 50 drawings are not reported. The LSSynduction motor, designed accordthe common stator. The rotor LS2rotor designed for vector control. LSSyR solution proposed here, withcapability.

C. Improvement of the pull-in capaThe pull-into step characteristics

examples are reported in Fig. 10. Thtotal inertia is reported for the three

The maximum inertia pulled-innominal torque (14.2 Nm) is 0.02inertia of the motor). At lower loalarger.

The state of the art SyR rotor Ldesigned with the criteria in [7] and Its pull-in capability is close to the motor LS1, at rated load. It is less respect to LS1, and then lower at lowat higher inertias.

Last, the proposed LS3 geometrimproved by 50% with respect to LWith respect to LS2, the cross sectiincreased as much as possible and, on top of the q axis has been remmore aluminum and reduce the d an

Figure 10. Pull-in torque characteristevaluated with trans

D. Comparison of the steady-state pTable I reports the steady state

LSSyR examples, at rated load. Tperformance of the benchmark IMthe improvements in efficiency. AFEA. The stator copper and rotor a

inum, to have the rotor is also consistent with the

SSyR motors had IM-like ection, then abandoned in flux barriers and all the ], such as the Synduction

evolution described by the otors have the same stator,

Hz IM competitor, whose S1 design is the like of a ding to [4] to comply with is a state of the art SyR The LS3 one is the final h a much increased pull-in

ability of the three LSSyR motor

he pull-in torque versus the machines. n by the LS1 motor at 20 kg m2 (four times the ads, the feasible inertia is

LS2, having four layers, is then filled with aluminum. one of the Synduction-like variable with inertia, with wer inertias and vice-versa

ry has a pull-in capability LS2 at all values of inertia. on of the four saliencies is most of all, the steel layer

moved for making room to d q rotor resistances.

tics of motors LS1 to LS3,

sient FEA.

performance performance of the three

he table includes also the M, for the quantification of All the results come from aluminum temperatures are

assumed to be 120 C, although the four motdifferent losses.

The steady state performance of the comparable to the one of the IM, meaning this a equal of a little higher and the powerresulting in a higher stator current (5.8 A vers

For what concerns LS2, the steady state significantly with respect to S1, as it can bethe much higher saliency and then torque tosynchronism. The final solution LS3 has a performance at steady-state, because it is noexactly for synchronous operation, but still mefficiency while increasing the pull-in capabil

TABLE 1: STEADY STATE PERFORMANCE OF THE EXANOMINAL LOAD

LS1 LS2

Line Voltage [V] rms 398 398

Phase Current [A] rms 5,8 4,79

Continuous Torque [Nm] 14,2 14,2

Rated Speed [rpm] 1500 1500

Output Power [W] 2231 2231

Rated Power Factor 0,718 0,763

Core loss [W] 61 48

Jouls loss, stator [W] 439 299

Joule loss, rotor [W] 121 31

Windage loss [W] 31 31

Efficiency 0,773 0,845

IV. EXPERIMENTAL RESULT

To validate the conclusion at section III, transient FEA are experimentally validated onprototype, whose rotor drawing is reportedmotor, originally equipped with permanent mintended for traction, is the one used in ratings are reported in Table II.

TABLE 2I: RATINGS OF THE PM-ASSISTED PROTOTYPBE MODIFIED FOR THE EXPERIMENTS AS L

Continuous Power kW 7

Peak Power kW 10

Base speed rpm 245

Max speed rpm 1000

Rated Current A 20

Rated Voltage V 257

Here the PMs have been removed from

tors have slightly

LS1 motor is hat the efficiency r factor is lower, sus 5.0 A). efficiency grows

e expected due to o Ampere ratio at little decrease in

o longer designed maintains a good lity significantly.

AMPLE MACHINES AT

LS3 Induction motor

398 398

4,95 5,0

14.2 15,1

1500 1381

2231 2183

0,745 0,794

50 50

318 330

48 168

31 31

0,834 0,790

S the results of the n a LSSyR motor

d in Fig.11. This magnets (PM) and [9], whose main

PE MOTOR PRIOR TO LSSYR

0

50

00

0

7

m the rotor and

copper bars have been inserted instethen welded at the hands to copperthe cross section of the rotor laminathe copper has been inserted. The tetwo stages. At first, the cage is onlyof each pole, as represented by the gLater on, the red copper bars have blayer, the smallest one. Fig. 12 repoprototype, in versions 1 and 2, beforthe small red bars.

Figure 11. Rotor lamination of the PM-to house a copper cage. Version 1 has o

version 2 has also the red bars

(a)

Figure 12. Rotor used for the experimVersion 2

A. Experimental setup

Fig. 13 exhibits the equipment usof the LSSyR prototype has been merotor by means of a known force apThe inertia of Version 1 is Jmot = 0.been assumed to have the same inerlittle impact when more rotating boare involved.

Figure 13. LSSyR motor prototype load

ead into the saliencies, and r end rings. Fig. 11 shows ations and the areas where ests have been divided into y in the three bigger layers grey copper bars in Fig. 11. been added into the fourth orts the picture of the rotor re and after the insertion of

-assisted SyR of [9], modified only the copper bars in grey, s in the outer layers.

(b)

mental tests. a) Verion 1; b)

ed for the tests. The inertia easured by accelerating the pplied with a known lever. .0057 kgm2. Version 2 has rtia, which is wrong but of odies other than the motor

ded by the DC machine.

The load torque is produced by a DC seAXEM series [10]. Its inertia is 0.0067 kgm2

of LSSyR prototype, DC machine and couplikgm2 (2,2 times the one of motor alone). Tincreased by including a disk of steel betweenleading to a total inertia of 0.021 kgm2. Theloaded with a resistor, producing a torque thto the speed. The resistance can be varied totorque curve. The LSSyR prototype is line 50Hz, compatible with the specifications of Tto a machine designed for being inverter drive

B. Pull-in capability of Versions 1 and 2 The pull-in characteristic of the LSSyR

evaluated, is reported in Fig. 14. The best exups are also reported, for the two versions. Ttests refer to only two values of inertia, thwithout the additional disk for extra inertia.has been varied progressively until the mocapable of synchronization. The difficucomprehensive tests stands in the effect ofcold conditions the pull-in torque is the one14. After a couple of successful start beginhowever, the rotor temperature is no longer tpull-in capability is reduced. If, for instancrepeat the same test many times for taking effect of voltage phase at time zero, thisbecause every time the temperature has chnecessary to wait until the machine (the rotorat room temperature again before running a ne

From the data related to Version 2, a gofound at the highest value of inertia, that givearound the rated torque value. At lower intorque is much lower than the FEA evaluatedhave some doubt about how comprehensive thas been.

The comparison between Versions 1 andbeneficial effect of the extra rotor bars. Themore than doubled.

Figure 14. Pull-in characteristic of the prototype mand 2, calculated with FEA and meas

C. Speed transients

The experimental and the FEA speed tranare reported in Fig. 15 for Version 1 and in Fi

ervomotor of the . The total inertia

ing is Jtot= 0.0124 he inertia can be n the two motors, e DC machine is

hat is proportional o change the load

started at 153V, Table II, that refer en.

prototype, FEA xperimental start-The experimental hat are with and The load torque

otor is no longer ulty of running f temperature. At e reported in Fig. nning from cold, the same, and the ce, one wants to into account the

s is not possible anged. It is then r, in particular) is ew test. ood agreement is es a pull-in torque nertia the pull-in d one, but still we the test campaign

d 2 confirms the e pull-in torque is

motor, in Versions 1 sured.

nsients at start-up ig. 16 for Version

2. Fig. 15 shows a good agreem

measures, and puts in evidence the ethe voltage vector. The load torquones reported in the characteristic of

Figure 15. Line start speed transient inertia is 12.4 10-3 kgm2 and Tload is 5.6Red lines: FEA calculated with differen

In Fig. 16 the two tests indicateFig. 14 are reported. In Fig. 16b, thpossible to see that the actual start to a sub synchronous speed and comin a second time. This never occurs w

(a)

(b)

Figure 16. Line start speed transient inertia is 12.4 10-3 kgm2 (a) and 20.9 10

(a) and 12.0 Nm (b). Black line: mcalculated.

V. CONCLUThe design and the performance o

Reluctance motors have been mo

ment between FEA and effect of the initial phase of e and total inertia are the f Fig. 14 (triangle marker).

with Version 1 rotor. Total

66 Nm. Black line: measured. nt initial voltage phase values.

ed with square markers in he one at high inertia, it is transient converges at first

mpletes the synchronization with FEA.

with Version 2 rotor. Total 0-3 kgm2 (b). Tload is 13.4 Nm

measured. Red lines: FEA .

SION of Line-Start, Synchronous

odeled and experimentally

verified. State of the art-SyR rotors show a better performance of LSSyR motors in the literature, and show to be competitive towards induction motor counterparts in terms of efficiency. The effect of filling as much space as possible with rotor conductors has been put in evidence, both in the model and experimentally. Future works will deal with the evaluation of the further optimization of the motor starting capability, by means of the choice of the number of layers and their placement.

REFERENCES [1] Honsinger, V.B.; , “Permanent Magnet Machines:

AsychronousOperation,”Power Apparatus and Systems, IEEE Transactions on, vol.PAS-99, no.4, pp.1503-1509, July 1980.

[2] Miller, T. J. E.; , “Synchronization of Line-Start Permanent-Magnet AC Motors,” Power Engineering Review, IEEE, vol.PER-4, no.7, pp.57-58, July 1984

[3] Kurihara, K.; Rahman, M.A.; , “High-efficiency line-start interior permanent-magnet synchronous motors,”IndustryApplications, IEEE Transactions on, vol.40, no.3, pp. 789- 796, May-June 2004.

[4] VB Honsinger; “Synchronous Reluctance Motor”, US Patent 3,652,885, March 1972

[5] Lawrenson, P.J.; Mathur, R.M.; , “Pull-in criterion for reluctance motors,”Electrical Engineers, Proceedings of the Institution of,vol.120,no.9, pp.982-986, September 1973

[6] Pettit, R.; Wagner, P.; Shaw, K.; , “Synchronous AC motors for process control over wide speed ranges ,” Textile Industry Technical Conference, 1988., IEEE 1988 Annual , vol., no., pp.6/1-6/8, 4-5 May 1988

[7] Vagati, A.; Franceschini, G.; Marongiu, I.; Troglia, G.P.; , “Design criteria of high performance synchronous reluctance motors ,”Industry Applications Society Annual Meeting, 1992., Conference Record of the 1992 IEEE, vol., no., pp.66-73 vol.1, 4-9 Oct 1992.

[8] Boroujeni, S.T.; Bianchi, N.; Alberti, L.; , "Fast Estimation of Line-Start Reluctance Machine Parameters by Finite Element Analysis," Energy Conversion, IEEE Transactions on , vol.26, no.1, pp.1-8, March 2011

[9] Pellegrino, G.; Armando, E.; Guglielmi, P., "Direct Flux Field-Oriented Control of IPM Drives With Variable DC Link in the Field-Weakening Region," Industry Applications, IEEE Transactions on , vol.45, no.5, pp.1619,1627, Sept.-oct. 2009

[10] Axem Servo Motors, http://www.parker.com/ [11] Magnet 7.3.0, by Infolytica Corporation. http://www.infolytica.com/