Synchronous Reluctance Motor Iron losses: Considering Machine Non-Linearity at MTPA, FW, and MTPV Operating Conditions

Transcription

1 Synchronous Reluctance Motor Iron losses: Considering Machine Non-Linearity at MTPA, FW, and MTPV Operating Conditions Hanafy Mahoud, Giacoo Bacco, Michele Degano, Nicola Bianchi and Chris Gerada Abstract Synchronous reluctance achine has high flux density fluctuations in the iron due to the high haronics results fro the rotor anisotropy. Thus, an accurate coputation of the iron losses is of paraount iportance, especially during the design stage. In this paper, a non-linear analytical odel considering the agnetic iron saturation and the slotting effect is proposed. The odel estiates accurately the iron losses at a wide range of operating speed. In addition, the accuracy of the non-linear odel when the achine is highly saturated, i.e. when it works along the MTPA trajectory, is presented and verified. The odel presented is general and can be applied to other configurations. A 36-slot four-pole achine, with three flux-barriers per pole is considered as a case study. Finite eleent analysis is used to validate the results achieved by eans of the non-linear analytical odel. Furtherore, an experiental setup is built to validate the siulation results. Index Ters Synchronous reluctance achine, Iron saturation, Non-linear analytical odels, Iron losses coputation, Finite eleent analysis. I. INTRODUCTION The iportance of deterining accurately the total iron losses in the whole otor structure is crucial even during the first design stage. In addition, the coputation tie of these losses is very iportant for reducing the diensioning phase in the product developent yields to a positive financial ipact. For that reason, the analytical odels are attractive tools for estiating these losses. Referring to the linear analytical odel presented in the literature [] [3], the over estiation of the iron losses results fro that linear odel leads to oversize the achine. Thus, it is necessary to find a fast and accurate analytical tool overcoe the assuptions of those linear odels. Fro the literature, it has been recognized that there are any approaches for coputing the iron losses of Synchronous reluctance (SynRel) achines. Most of the are based on the analysis, as in [4] [8], or on analytical H. Mahoud is with the Departent of Electrical and Electronic Engineering, University of Nottingha, NG7RD, Nottingha (UK), and also with the Departent of Electrical Power and Machines, Cairo University, 63 Giza, Egypt (e-ail: eng.hanafy4@yahoo.co, Hanafy.Mahoud@nottingha.ac.uk). G. Bacco and N. Bianchi are with the Departent of Industrial Engineering, University of Padova, 353 Padova, Italy (e-ail: giacoo.bacco@phd.unipd.it), (e-ail: nicola.bianchi@unipd.it; nicola.bianchi@dii.unipd.it). M. Degano and C. Gerada are with the Departent of Electrical and Electronic Engineering, University of Nottingha, NG7RD, Nottingha (UK), and also with the University of Nottingha Ningbo China, Ningbo, China. (e-ail: ichele.degano@nottingha.ac.uk), (e-ail: chris.gerada@nottingha.ac.uk) analysis for specific stator MMF and rotor geoetry, as in [], [3]. The analytical odel presented in [] considers all haronics of the stator MMF and is valid for a general rotor geoetry. However, it focuses only on the eddy current loss coputation in the stator teeth. Besides, the rotor end angles are optiized in order to iniize those losses disregarding their effect on the torque ripple, however it has a high ipact on reducing both stator teeth and yoke iron losses, as well [9]. The previous analysis has been carried out neglecting both: (a) the stator slotting effect, and (b) the agnetic iron saturation effect. These phenoena strongly affect the air-gap flux density distribution [] []. In fact, soe dips appear in the air-gap flux density distribution due to the agnetic reluctance variations in the air-gap caused by the slot openings [3] [6]. In addition, the agnetic voltage drops, occurring in the stator and rotor iron, liit the agnetic fluxes flowing through the iron parts of the otor. As a consequence, the flux density variations in these regions are reduced [], [7]. It has been observed that the iron losses coputed in [9] result to be overestiated, especially if they are coputed along the MTPA trajectory. To fill this gap, the ai of this paper is to include the achine non-linearities in the analytical odel so as to copute ore accurately the iron losses in various iron parts. An novel aspect introduced in this work, with respect to the presented stator yoke stator tooth stator slot shaft 3 rd barrier nd barrier st barrier channel q - axis nd island st island Fig. : Sketch of the SynRel achine under study. d - axis 3 rd island

2 odel in [], is the extension to the iron losses prediction in the various iron parts of the achine, not only in the stator teeth. The hysteresis losses are considered for the fundaental coponent, because of their significant aplitude, especially if the achine is working along the MTPA trajectory. Fig. reports a D sketch of SynRel achine showing the different iron parts of the stator, i.e. teeth and yoke, and the rotor, i.e., islands and channels. Once again, a 36-slot fourpole achine with three flux-barriers per pole is used, as an exaple, however, the non-linear analytical odel is general for a different nuber of poles, stator slots, flux-barriers, as can be recognized in [8], [9]. The design specifications, of the studied SynRel, are suarized in Table I. TABLE I: Geoetrical data of SynRel otor used in section III. Inner stator Diaeter D o Inner stator Diaeter D 5 Stack length L stk 4 Nuber of stator slots Q s 36 Nuber of stator pole pairs p s Air-gap length g.35 First flux barrier ends angle θ b 8 Second flux barrier ends angle θ b 5.4 Third flux barrier ends angle θ b The paper is organized as follows: Section II briefly describes the non-linear analytical odel. Section III reports the iron losses when the otor operate along the MTPA, FW, and MTPV trajectory, respectively. Furtherore, it shows the coparison between the linear and non-linear analyses results for all the above conditions. Section IV studies the ipact of the local saturation close to the iron ribs on the non-linear analytical odel accuracy. siulations and an experiental easureents validate the results obtained by eans of the analytical analyses. II. NON-LINEAR ANALYTICAL MODEL In this section, the analytical odel considers the sae pereance function discussed in [], [5]. Then, the air-gap flux density is updated. Its distribution is confired by the analysis, as shown in Fig.. Air-gap flux density [T] ϑ s [Elec. Deg.] Fig. : Air-gap flux density distribution at J = A/ and αi e = 8, i.e., along the MTPV trajectory. The agnetic voltage drop in the iron paths can be considered by eans of an additional equivalent voltage drop in the air-gap. In other words, the air-gap length is increased by a specific factor: the saturation factor K sat [], [6], []. Due to the coplex rotor structure, it is difficult to consider the agnetic saturation in both stator and rotor iron by eans of a unique equivalent factor. l ht wt ψ y(3) Ds De ψ y(9) ψ y(8) ψ y(7) ψ y(6) B t3 ψ y(5) ψ y(4) Fig. 3: The ain idea behind the ghost paths. ψ y(3) ψ y() ψ y() As shown in Fig. 3, the ghost flux line enters fro the tooth n and exits fro the tooth (n + Q s /p). It is equivalent to consider that every flux line travels for a full pole pitch. Referring to the stator, Fig. 3 shows that the agnetic voltage drop in each flux path ghost path occurs in the yoke and two teeth. Considering half of the path, the agnetic drop can be presented by K sattn in front to the n th tooth. It can be expressed as K sattn B y3 = + ψ t n + ψ ytn gh gn, n =,..., Q s /p () where ψ tn is the agnetic voltage drop in the n th tooth. ψ ytn is the agnetic drop which occurs in half the flux path in the stator yoke. The subscript yt indicates that the yoke drop is considered as an additional agnetic drop along the tooth. H gn is the field intensity in front of the n th tooth. The detailed coputations of both ψ tn and ψ ytn will be discussed later on. On the other hand, the voltage drop in the rotor islands and channels can be considered by adopting K satis,i to increase the air-gap in front of each rotor iron part. It can be coputed as K satis,i = + µ ψ isi gb gis,i, i =,..., N b + () where ψ isi is the agnetic voltage drop in the i th rotor island. B gis,i is the air-gap flux density in front to the i th island. Indeed, the coputation for the rotor channel is carried out when i = N b +. A. Magnetic voltage drop in the stator iron ) Stator teeth: The flux density variation in the n th stator tooth can be coputed as B tn ( ) = D w t γs+nα slot γ s+(n )α slot B g (ϑ s, ) dϑ s (3)

3 Then, fro the actual iron B-H curve, shown in Fig. 4, the field intensity is deterined. Hence, ψ tn can be coputed as B tn (t) H tn (t) ψ tn = H tn (t) h t (4) where h t is the height of the stator tooth. B [T] H [A/] x 5 Fig. 4: Magnetization curve of the ferroagnetic aterial used. ) Stator yoke: Starting fro a generic tooth, the actual flux flowing through the s th section of the stator yoke is coputed as φ ys n=s = w t L stk B tn w tl stk Q s /p n= s=q s/p s= n=s B tn (5) n= Consequently, the flux density, field intensity, and agnetic voltage drop in the s th section can be achieved as B ys (t) = φ y s (t) h y L stk H ys (t) ψ ys = H ys (t) l (6) where l = π(d e h y )/Q s is the average length of one backiron sector. The ghost flux line encounters Q s /p agnetic voltage drops along its path. These agnetic voltage drops can be sued, with the sign of the corresponding flux in the sae sector. At the end, the ghost flux line has gathered the sae overall agnetic voltage drop of the actual flux line of that particular tooth and can be expressed as n+ Qs p ψ ytn = ψ y [K] sign φ y [K], n =,..., Q s/p K=n (7) The occurs since half the path of the line is considered. As an exaple, the solid path shown in Fig. 3 starts fro tooth n = (below the first slot) and it ends in tooth n =. The total agnetic voltage drop associated to this line is ψ yt = ψ y + ψ y ψ y9 (8) On the other hand, the dotted path shown in Fig. 3 sees ψ yt5 = ψ y ψ y9 ψ y... ψ y3 = ψy5 (9) where the last equivalence is the ain assuption of this coputation. In fact it is like iplying that ψ y6 is equal and opposite to ψ y3, siilarly, ψ y7 to ψ y and so on. All ters are canceled out but the ter ψ y5. B. Magnetic voltage drop in the rotor iron Fig. 5 shows the actual flux paths in the different rotor iron parts, such as the flux enters in and exits fro each rotor island and channel. The flux flowing through the i th flux barrier is coputed as φ isi ( ) = φ ini ( ) φ b i ( ) + φ bi ( ) where φ bi ( ) is given by () φ bi ( ) = φ ini ( ) φ outi ( ) () By integrating the air-gap flux density, φ ini ( ) and φ outi ( ) are coputed as φ ini ( ) = φ outi ( ) = π p ϑ b i π p ϑ b i π p +ϑ b i B g (ϑ s, ) DL stk B g (ϑ s, ) DL stk π p +ϑ b i dϑ s () dϑ s (3) On the other hand, the flux flowing through the rotor channel is coputed as φ ch = φ chi φ cho (4) where φ chi and φ cho are derived by integrating the air-gap flux density as follows φ chi = φ cho = π p ϑ b Nb + π p +ϑ B g (ϑ s, t) DL stk B g (ϑ s, t) DL stk π p +ϑ b Nb + dϑ s (5) dϑ s (6) The aforeentioned island and channel fluxes are the d axis fluxes. The q axis flux for the islands and the channels is is is3 chi ch in3 b3 in b in b out out b out3 cho q-axis b3 d-axis Fig. 5: Actual flux paths in the different rotor iron parts.

4 are also coputed through the following approxiation φ isq = φ b,..., φ iqn b = φ b N b + φ b N b (7) φ chq = φ b N b The island q axis flux equations represent the average between the upper and lower barrier fluxes. In order to get the flux density in each i th island, all these fluxes are referred to the iddle section of the iron segent. For each i th island, the width w ri is known fro the drawing or coputed as in [], while the length is approxiated through therefore B isi = l isi = D ϑ bi + ϑ bi B isdi = φ isd i w ri L stk, B isqi = φ isq i l isi L stk B chd = φ chd, B chq = φ chq w ch L stk l ch L stk B isd i + B isq i, B ch = B chd + B chq (8) (9) where w ch = ( kair )D sin ( π p ϑ bn b ), lch = (D re +D ri )/, k air is the ratio between the su of the barriers thicknesses and the total length of the rotor along the q axis. C. Iterative approach The saturation factors reported in () and () are not fixed but they are iteratively coputed so as to adjust the agnetic odel paraeters. The nuber of the iterations depends on the saturation level in the achine iron parts. As an exaple, Fig. 6 explains the iterative approach for coputing K sattn in front of the n th tooth. III. ANALYSIS RESULTS AND VALIDATION In this section, the results of the linear odel at the operating points B (along MTPA trajectory) and B (along FW trajectory) are copared by the results of the non-linear odel. The current density is the sae for both operating conditions (J = 3A/ ) and the current vector angles are 45 and 8, respectively. For the sake of studying the odel accuracy at high saturation level, the coparison is repeated at current density J = 6A/. Furtherore, the non-linear odel is confired by the analysis in various operating points, along the MTPA trajectory, under the FW operation, and along the MTPV trajectory, i.e., at J = A/ and α e i = 8. Fro [], [9], it is recognized that the rotor losses are negligible coparing to the stator losses and can be neglected if the rotor is lainated and not rotating at very high speed. Therefore, the ain focus of this paper is on the stator iron losses coputation. A. MTPA operating condition Referring to the operating point B, Fig. 7 and Fig. 8 show the flux density variations in the stator teeth and yoke resulting K sat = K sat-prev. start read B g( θ s ) and actual iron B-H curve. define K sat-step =., Intial value of K sat =, Intial value of K sat-prev. =. x = copute K sat as in () and tn K sat-var = K sattn / K sat no x = if K sat-var >? end yes Iprove the air-gap flux density in front of the n-th stator tooth if x =? yes save K sattn = K sat Fig. 6: Iterative approach for coputing the accurate saturation factor in front to the n th stator tooth. fro the non-linear slotted odel. Three teeth are considered since there are three slots per pole and per phase. It is noted that there is a satisfactory agreeent between the iproved odel and the analysis. Table II shows the iron losses in the various iron parts of the stator core. The eddy current losses (due to all haronics) and the hysteresis loss (due to the fundaental haronic only) are reported. Again, the nonlinear analytical odel results are validated by eans of analysis, as shown in Table II. Coparing Table II with the corresponding Table III, which reports the iron losses resulting fro the linear analytical odel, it is noted that the iron losses are overestiated, by 9.5% with respect to those predicted by. This coparison rearks the accuracy of the non-linear analytical odel (axiu error approaches %) even if the local saturation close to the rotor iron ribs is disregarded. B. FW operating condition The non-linear analytical and odel results are copared together at the operating point B, as shown in Fig. 9, Fig., and Table IV. In FW conditions, there is a great reduction of the ain flux. As a consequence, the saturation level of the achine is low. The local saturation at the iron ribs is reduced too. This increases the accuracy of the non-linear no K sat = K sat-prev. K sat = K sat-prev. K sat-step

5 - Toothy Toothy ϑ [Elec.yDeg.] - (a) odel Toothy Toothy ϑ [Elec.yDeg.] (b) odel Fig. 7: Flux density variations in stator teeth at operating Point B resulting fro the non-linear odels. Yoketfluxtdensityt[T] ϑ [Elec.tDeg.] Fig. 8: Flux density variation in stator yoke at operating Point B resulting fro the non-linear odels. odel (axiu error.8%). Additionally, it is noted that, in point B the haronic losses are doinant, while the losses associated to the fundaental are quite low. This is due to both the relative increase of haronic content and the frequency. Siilarly, coparing Table IV to the corresponding results of the linear analytical odel, it is noted that the linear odel overestiates the coputed losses by 8.5%. Fro Table II and Table IV, it can be noted that the neglecting of the iron ribs becoes ore realistic at the FW operating condition and the odel error is reduced fro.5% (at point B) to.5% (at point B ). C. MTPV operating condition Hereafter, the otor is supplied with J = A/ and αi e = 8. It works at operation point along the MTPV TABLE II: Eddy current and hysteresis loss densities of the stator teeth and yoke at the operating point B resulting fro the non-linear analytical and odels. Iron part al h = h > h = h > Yoke First tooth Second tooth Third tooth Yoke First tooth Second tooth Third tooth al Stator tooth.4. Stator yoke Model error % TABLE III: Eddy current and hysteresis loss densities of the stator teeth and yoke at the operating point B resulting fro the linear analytical and odels. Iron part al h = h > h = h > Yoke First tooth Second tooth Third tooth Yoke First tooth Second tooth 9.84 Third tooth.5.3 al Stator tooth 3..7 Stator yoke Model error < % trajectory (point C). Fig. and Fig. show the flux density variations in the stator teeth and yoke estiated by the nonlinear analytical and odels. Hence, the stator core iron losses are estiated by both non-linear odels and reported in Table V. There is a good agreeent between both non-linear odels. The non-linear analytical odel overestiates the odel predicted losses by.5% only. When the otor runs along the MTPV operating trajectory, the current aplitude is reduced. Thus, the d axis current is reduced, and hence the ain flux of the achine is decreased ore than in the previous FW operating condition, as shown in Fig. and Fig.. In [], it is entioned that neglecting the iron saturation applied to the linear analytical odel is well suited to the MTPV operations due to low current density at this operating region. The linear analytical and odels are applied to the

6 Toothy Toothy [Elec.yDeg.] (a) odel Toothy Toothy [Elec.yDeg.] (b) odel Fig. 9: Flux density variations in stator teeth at operating Point B resulting fro the non-linear odels. Yoketfluxtdensityt[T] ϑ [Elec.tDeg.] Fig. : Flux density variation in stator yoke at operating Point B resulting fro the non-linear odels. sae operating point. The estiated iron losses are copared to those reported in Table V, hence, it is recognized that the iron losses resulting fro the linear odel are overestiated by 4.5%. To understand the effect of neglecting the iron saturation on the analytical odel accuracy, a sei non-linear odel conserving the stator slotting effect only is applied at the sae operating condition. It is recognized that this analytical odel overestiates the non-linear odel by.3%. IV. ACCURACY ROBUSTNESS OF THE NON-LINEAR ANALYTICAL MODEL The non-linear analytical odel does not consider the local saturation close to the rotor iron ribs. This local saturation interacts with the stator teeth and causes the different haronic contribution. This section ais to study the effect of this assuption on the accuracy of the iron losses coputations when the otor is highly saturated, i.e., it works along the MTPA TABLE IV: Eddy current and hysteresis loss densities of the stator teeth and yoke at the operating point B resulting fro the non-linear analytical and odels..5 Iron part al h = h > h = h > Yoke First tooth Second tooth Third tooth Yoke.6.86 First tooth.4.4 Second tooth.5.75 Third tooth al Stator tooth.8.7 Stator yoke Model error.5% Toothy Toothy e [Elec.yDeg.] (a) odel Toothy Toothy e [Elec.yDeg.] (b) odel Fig. : Flux density variations in stator teeth resulting fro the non-linear odels at the operating point C. trajectory. The accuracy of the non-linear analytical odel for coputing the other achine perforance paraeters at high saturation level has been checked in [], [6]. In this section, the odel accuracy is checked at J = 6 A/ and α e i = 59. Table VI reports the iron losses in the stator core at the aforeentioned operating condition. It is noted that, the non linear analytical odel overestiates the non-linear slotted odel by about %, uch better than the linear analytical odel which overestiates the losses by about 5% in this condition. In addition, Fig. and Fig. show a satisfactory agreeent between the non-

7 Yoketfluxtdensityt[T] e [Elec.tDeg.] Fig. : Flux density variation in stator yoke resulting fro the non-linear odels at the operating point C. TABLE V: Eddy current and hysteresis loss densities of the stator teeth and yoke resulting fro the non-linear odels at the operating point C. Iron part al h = h > h = h > Yoke First tooth Second tooth Third tooth Yoke.35.4 First tooth.7.3 Second tooth Third tooth.6.7 al Stator tooth Stator yoke.63.7 Model error.5% - Toothy Toothy [Elec.yDeg.] - (a) odel [Elec.yDeg.] (b) odel Toothy Toothy Fig. 3: Flux density variations in stator teeth at J = 6A/ and αi e = 59 resulting fro the non-linear odels. Yoketfluxtdensityt[T] - linear analytical and odels. TABLE VI: Eddy current and hysteresis loss densities of the stator teeth and yoke at J = 6A/ and αi e = 59 resulting fro the non-linear odels. Iron part al h = h > h = h > Yoke First tooth Second tooth Third tooth Yoke First tooth Second tooth Third tooth al Stator tooth Stator yoke.6.5 Model error % Finally, an experiental validation is carried out by validating the output power/torque ap of both the siulation ϑ [Elec.tDeg.] Fig. 4: Flux density variation in stator yoke at J = 6A/ and α e i = 59 resulting fro the non-linear odels. and the experiental easureents. This can be carried by fixing the input power for both the siulation and the actual otor under the experiental test. An indication of an accurate losses coputation can be achieved if the output power of both siulation and test are atched, for the sae input power. The experiental validation of the output torque ap is shown in Fig. 5. It is noted that there is a satisfactory agreeent between siulation results and experiental easureents. V. CONCLUSIONS This paper describes a rapid non-linear analytical odel for coputing the iron losses in SynRel achine. This odel is well suited to copute the flux density variation in the various iron parts of the SynRel achine. It allows the iron losses to be coputed in any iron part. The non-linear odel is checked in a wide operating speed range. Thus, the coputations are carried out at the operating points B (MTPA),

8 Peak q-axis current [A] analytical Peak d-axis current [A] experiental Fig. 5: Torque ap results fro the experiental easureents and the siulation. B (FW), and C (MTPV). The achine speed, saturation level, and haronic contribution are copletely different in these operating conditions. It is noted that the non-linear odel exhibits a good agreeent with the analysis. The results of the non-linear odel are also copared with the linear odel so as to evaluate the iproveent occurred in the iron losses coputations. It is noted that the error is reduced fro 9.5% to.5%, fro 8.5% to.5%, and fro 4.5% to.5%, at the operating points B, B, and C, respectively. Although the current aplitude is reduced at the operating point C (MTPV), the iron agnetic saturation affects the result accuracy. This can be noted fro the results of the linear slotted analytical odel result. The error of this sei nonlinear odel is.3% which is, of course, higher than that of the non-linear analytical odel. In addition, the non-linear odel is checked when the achine is saturated at current density 6A/ along the MTPA trajectory. Even if the saturation level is doubled, the non-linear odel results are satisfactory, yielding a slight overestiation of the losses around % with respect the coputation. This overestiation could be useful for the sake of safety for considering the anufacturing process effects on the agnetic properties of the achine iron. The non-linear analytical odel is a quick tool to estiate the iron losses of the achine. For one electric period, the siulation tie is about seconds. This advantage of the odel is very iportant for reducing the diensioning phase in the otor developent yields to a positive financial ipact. RERENCES [] M. Barcaro and N. Bianchi, Air-gap flux density distortion and iron losses in anisotropic synchronous otors, in IEEE Transactions on Magnetics, vol. 46, no., pp. 6, Jan. [] L. Ma, M. Sanada, S. Morioto, and Y. Takeda, Iron loss prediction considering the rotational field and flux density haronics in IPMSM and SynRM, IEE Proceedings - Electric Power Applications, vol. 5, no. 6, pp , Nov 3. [3] G. Pellegrino, P. Guglieli, A. Vagati, and F. Villata, Core losses and torque ripple in IPM achines: Dedicated odeling and design tradeoff, in IEEE Transactions on Industry Applications, vol. 46, no. 6, pp , Nov. [4] M. J. Kaper, F. S. V. der Merwe, and S. Williason, Direct finite eleent design optiisation of the cageless reluctance synchronous achine, IEEE Transactions on Energy Conversion, vol., no. 3, pp , Sep 996. [5] J. H. Lee, I. K. Lee, Y. H. Cho, and T. W. Yun, Characteristics analysis and optiu design of anisotropy rotor synchronous reluctance otor using coupled finite eleent ethod and response surface ethodology, IEEE Transactions on Magnetics, vol. 45, no., pp , Oct 9. [6] P. Hudak, V. Hrabovcova, P. Rafajdus, and J. Mihok, Core loss analysis of the reluctance synchronous otor with barrier rotors, Journal of Electrical Engineering, vol. 55, pp , 4. [7] F. Leonardi, T. Matsuo, and T. A. Lipo, Iron loss calculation for synchronous reluctance achines, in Proceedings of International Conference on Power Electronics, Drives and Energy Systes for Industrial Growth, vol., Jan 996, pp vol.. [8] T. Rainosoa, I. Rasoanarivo, and F. M. Sargos, Reluctance network analysis of high power synchronous reluctance otor with saturation and iron losses considerations, in 6 th International Power Electronics and Motion Control Conference, Portoroz, Slovenia, Aug 6, pp [9] H. Mahoud, M. Degano, G. Bacco, N. Bianchi, and C. Gerada, Synchronous reluctance otor iron losses: al odel and optiization, In IEEE Energy Conversion Congress and Exposition (ECCE 8), in press. [] P. Akiki, M. H. Hassan, J. C. Vannier, M. Bensetti, B. DagusÃ, D. Prieto, and M. McClelland, Non-linear analytical odel for a ultiv-shape IPM with concentrated winding, in 6 XXII International Conference on Electrical Machines (ICEM), Lausanne, Switzerland, Sept 6, pp [] H. Mahoud, N. Chiodetto, and N. Bianchi, Magnetic field analytical coputation in synchronous reluctance achines considering the iron saturation, In IEEE Energy Conversion Congress and Exposition (ECCE 6), Milwaukee, WI, USA, 8- Sept, 6. [] D. Prieto, P. Dessante, J. C. Vannier, B. DagusÃ, X. Jannot, and J. Saint-Michel, Multi-physic analytical odel for a saturated peranent agnet assisted synchronous reluctance otor, IET Electric Power Applications, vol., no. 5, pp , 6. [3] Z. Zhu and D. Howe, Instantaneous agnetic field distribution in brushless peranent agnet dc otors. iii. effect of stator slotting, IEEE Transactions on Magnetics, vol. 9, no., pp. 43 5, Jan 993. [4] J. Gieras, al approach to cogging torque calculation of PM brushless otors, IEEE Transactions on Industry Applications, vol. 4, no. 5, pp. 3 36, Sept 4. [5] D. Zarko, D. Ban, and T. Lipo, al calculation of agnetic field distribution in the slotted air gap of a surface peranent-agnet otor using coplex relative air-gap pereance, IEEE Transactions on Magnetics, vol. 4, no. 7, pp , July 6. [6] H. Mahoud and N. Bianchi, Nonlinear analytical odel of eccentric synchronous reluctance achines considering the iron saturation and slotting effect, IEEE Transactions on Industry Applications, vol. 53, no. 3, pp. 7 5, May 7. [7] M. Ojaghi and S. Nasiri, Modeling eccentric squirrel-cage induction otors with slotting effect and saturable teeth reluctances, IEEE Transactions on Energy Conversion, vol. 9, no. 3, pp , Sept 4. [8] H. Mahoud and N. Bianchi, Eccentricity in synchronous reluctance otors;part I: al and finite-eleent odels, in IEEE Transactions on Energy Conversion, vol. 3, no., pp , June 5. [9] G. Bacco and N. Bianchi, Choice of flux-barriers position in synchronous reluctance achines, in 7 IEEE Energy Conversion Congress and Exposition (ECCE), Oct 7, pp [] K. Abbaszadeh and F. R. Ala, On-load field coponent separation in surface-ounted peranent-agnet otors using an iproved conforal apping ethod, in IEEE Transactions on Magnetics, vol. 5, no., pp., Feb 6. [] N. Bianchi, H. Mahoud, and S. Bolognani, Fast synthesis of peranent agnet assisted synchronous reluctance otors, in IET Electric Power Applications, vol., no. 5, pp. 3 38, 6.

9 Hanafy Mahoud received the bachelor s degree and the M.Sc. degree in electrical engineering in 9 and respectively, fro Cairo University, Cairo, Egypt. During his M.Sc. period he worked on faults detection and perforance analysis of the induction achines. He received his a PhD degree fro Padova University, Italy. Now, he is a Research Fellow at the PEMC group, Departent of Electrical and Electronic Engineering, Nottingha University, UK. His research activities deals with the analysis of synchronous peranent agnet and reluctance achines, focusing on analytical and finite eleent odeling, faulty conditions analyses. Besides, his current research interests are high efficiency achine design for the advanced propulsion systes, as well as analytical odeling of various electrical achines. Giacoo Bacco (s 5) received both the BS in Energy Engineering in 3 and the MS in Electrical Engineering in 6 fro the University of Padova, Padova, Italy. Currently he is a PhD student of the Departent of Industrial Engineering at the sae University. His research interests are focused in the analysis and design of synchronous achines, in particular synchronous reluctance (REL) otors. Michele Degano (M 5) received the Laurea degree in electrical engineering fro the University of Trieste, Trieste, Italy, in, and the Ph.D. degree in industrial engineering fro the University of Padova, Padova, Italy, in 5. During his doctoral studies, he cooperated with several local copanies for the design of peranent-agnet achines. In 5, he joined the Power Electronics, Machines and Control Group, The University of Nottingha, Nottingha, U.K., as a Research Fellow, where he is currently an Assistant Professor teaching advanced courses on electrical achines. His ain research interests include design and optiization of peranent-agnet achines, reluctance and peranent-agnet-assisted synchronous reluctance otors through genetic optiization techniques, for autootive and aerospace applications, ranging fro sall to large power. Nicola Bianchi (M 98 SM 9 F 4) received the M.Sc. and Ph.D. degrees in electrical engineering fro the University of Padova, Padua, Italy, in 99 and 995, respectively. In 998, he joined the Departent of Electrical Engineering, University of Padova, as an Assistant Professor. Since 5, he has been an Associate Professor of electrical achines, converters, and drives. His activity is conducted in the Electric Drive Laboratory, Departent of Electrical Engineering, University of Padova. His teaching activity deals with the design ethods of electrical achines, where he introduced the finite-eleent analysis of achines. He is the author or coauthor of several scientific papers and international books on electrical achines and drives. He is recipient of five awards for best conference and journal papers His research activity is in the field of design of electrical achines, particularly for drive applications, in which he is responsible for various projects for local and foreign industries. He is an IEEE Fellow eber and a eber of the Electric Machines Coittee and the Electrical Drives Coittee of the IEEE Industry Applications Society. He served as technical progra chair for IEEE ECCE 4 and is Associate Editor of IEEE Trans on IA and IET-EPA proceedings. Chris Gerada (M 5 SM ) received the Ph.D. degree in nuerical odelling of electrical achines fro The University of Nottingha, Nottingha, U.K., in 5. He was a Researcher with The University of Nottingha, working on highperforance electrical drives and on the design and odelling of electroagnetic actuators for aerospace applications. Since 6, he has been the Project Manager of the GE Aviation Strategic Partnership. In 8, he becae a Lecturer in electrical achines, in, as an Associate Professor, and in 3, a Professor at The University of Nottingha. His ain research interests include the design and odelling of high-perforance electric drives and achines. Prof. Gerada serves as an Associate Editor for the IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS and is the Chair of the IEEE Industrial Electronics Society Electrical Machines Coittee.