Performance Comparison of Conventional Synchronous ... - MDPI

energies Article

Performance Comparison of Conventional Synchronous Reluctance Machines and PM-Assisted Types with Combined Star–Delta Winding Mohamed Nabil Fathy Ibrahim 1,2, *, Essam Rashad 3 and Peter Sergeant 1,4 1 2 3 4

*

ID

Department of Electrical Energy, Metals, Mechanical Constructions and Systems, Ghent University, 9000 Ghent, Belgium; [email protected] Electrical Engineering Department, Kafrelshiekh University, Kafr El-Sheikh 33511, Egypt Electrical Power and Machines Department, Tanta University, Tanta 31527, Egypt; [email protected] Flanders Make, The Strategic Research Center for the Manufacturing Industry, B-8500 Kortrijk, Belgium Correspondence: [email protected] or [email protected]; Tel.: +32-468262801

Received: 23 August 2017; Accepted: 20 September 2017; Published: 27 September 2017

Abstract: This paper compares four prototype Synchronous Reluctance Motors (SynRMs) having an identical geometry of iron lamination stacks in the stator and rotor. Two different stator winding layouts are employed: a conventional three-phase star connection and a combined star–delta winding. In addition, two rotors are considered: a conventional rotor without magnets and a rotor with ferrite magnets. The performance of the four SynRMs is evaluated using a two-dimensional (2D) Finite Element Model (FEM). For the same copper volume and current, the combined star–delta-connected stator with Permanent Magnets (PMs) in the rotor corresponds to an approximately 22% increase in the output torque at rated current and speed compared to the conventional machine. This improvement is mainly thanks to adding ferrite PMs in the rotor as well as to the improved winding factor of the combined star–delta winding. The torque gain increases up to 150% for low current. Moreover, the rated efficiency is 93.60% compared to 92.10% for the conventional machine. On the other hand, the impact on the power factor and losses of SynRM when using the star–delta windings instead of the star windings is merely negligible. The theoretical results are experimentally validated using four identical prototype machines with identical lamination stacks but different rotors and winding layouts. Keywords: combined star–delta winding; design; FEM; PM-assisted; synchronous reluctance motor (SynRM)

1. Introduction In the literature, several types of Synchronous alternating current (AC) machines can be found, e.g., Interior and Surface Permanent Magnet Motors (IPMs and SPMs) and Synchronous Reluctance Motors (SynRMs) [1–5]. Thanks to their high efficiency, synchronous machines have received great interest in several applications, e.g., electric vehicles and photovoltaic (PV) pumping systems [3–5]. Recently, more research focus has been given to the SynRMs. This is thanks to their low cost and high efficiency compared to induction machines. In addition, a rotor of several flux-barriers per pole is always employed that has a simple and a rugged structure. The rotor losses are low, and hence it can work properly at higher temperatures [6–8]. However, SynRMs have two main disadvantages: the high torque ripple and the low power factor [3–8]. On the one hand, the high torque ripple can be reduced by two main approaches: selecting optimal geometrical parameters for the rotor flux-barriers (in particular, the flux-barrier angles) and skewing the rotor with respect to the stator [3,6]. These two methods can be combined together, resulting in a SynRM design with a torque ripple of Energies 2017, 10, 1500; doi:10.3390/en10101500

www.mdpi.com/journal/energies

Energies 2017, 10, 1500

2 of 18

less than 10% [3]. On the other hand, the rather poor power factor of SynRMs requires a high kVA inverter. This means that the low cost of SynRM may be compensated for by a higher cost inverter [3,8]. In order to improve the power factor and to enhance the torque density and efficiency of SynRMs, permanent magnets (PMs) are inserted into the rotor flux-barriers, resulting in a PM-assisted SynRM (PMaSynRM) [9]. Ferrite PMs are always employed in a PMaSynRM to reduce the machine’s cost compared to conventional permanent magnets synchronous machines (PMSMs) [9,10]. In addition, they can work at higher temperatures compared to PMSMs with rare-earth magnets. This indeed increases the reliability of PMaSynRMs. Moreover, the stator winding configuration can be a possible way to further improve the overall SynRM performance [11,12]. In the literature, much research work on SynRMs and PMaSynRMs can be found [3–16]. For example, in [8], SynRM performance is compared for different electrical steel grades. It is shown that the electrical steel grade has an enormous influence on SynRM efficiency: about 9% points higher for NO20 compared to M600-100A. In addition, the output torque increases by about 8%. However, there is a negligible impact on the SynRM’s power factor for different steel grades. In [9], an experimental investigation on PMaSynRM with ferrite magnets for automotive applications is presented. In addition, the irreversible demagnetization of ferrite magnets and mechanical strength are considered. A dual-phase material is utilized in the SynRM rotor design for traction applications in [14]. This is done by using a non-magnetic material in the radial and tangential ribs of the flux-barrier, leading to an increased saliency ratio. Eventually, the overall machine performance is improved compared to the conventional rotor design. A design and optimization of a high speed PMaSynRM for traction applications is investigated in [15]. The study takes into account both highways and city driving cycles. Various experimental tests on SynRM and PMaSynRM are presented in [16]. It is shown that inserting PMs in the rotor leads to a 10% increase in the SynRM’s torque at low speed and a 50% increase in a field weakening operation. The influence of rotor skewing is studied as well, showing a decrease in the torque ripple to about one third. However, the machine’s torque is slightly decreased. Moreover, it is evident that the SynRM’s power factor is improved in the overall operating regions when PMs are inserted in the rotor. Overlapping fractional slot concentrated windings are applied to a SynRM in [11]. It is shown that this winding type increases the power density and efficiency. In addition, it increases the robustness and the thermal behaviour of the SynRM. However, several challenges still need to be addressed in the literature for further research for this type of winding, e.g., high torque ripple and iron loss due to high magnetomotive force (MMF) space harmonics. In addition, the power factor is too poor. A combined star–delta winding is applied to a SynRM and compared to the conventional star connection in [12]. It is found that the SynRM’s output torque increased by about 5% at the rated conditions compared to the conventional star connection. This is because of the improved winding factor of the star–delta connection. In addition, the efficiency of the SynRM was slightly increased with a star–delta connection. However, there is no influence on the power factor using the different windings. The work presented herein compares the performance of four prototype SynRMs having an identical geometry of iron lamination stacks in the stator and rotor; two rotors (with and without PMs) and two stator winding connections (star and combined star–delta) are considered. The prototype of a combined star–delta winding in the stator and ferrite PMs in the rotor could be a very promising candidate for several industrial applications, e.g., PV pumping applications and electric vehicle traction. 2. Prototype SynRM Design In this section, the design of four machines with identical geometry but two different rotors and stator winding topologies is given. The optimization of the machines is not the goal of this paper: it is done in [17,18] and only briefly presented here. In this study, a 36 slot and 4 pole SynRM is employed with the geometrical parameters listed in Table 1. Two distributed winding configurations are used: the first configuration is the conventional

Energies 2017, 10, 1500

3 of 18

star-connected winding Energies 2017, 10, 1500

and the second one is two three-phase winding sets connected in a combined 3 of 17 star–delta connection. Both of the winding configurations result in three phases as shown in Figure 1. 2. As shown in the literature [19–21], connection 26, with a conductor cross-section area of 1.573 mm This meansisthat the number of stator slots/poles/phases (q) is 3. The number of turns per slot 2 . As two possible connections of star delta coilscross-section can be made:area either two coils areshown connected in of the star connection is 26, withand a conductor of the 1.573 mm in the series or they are connected in parallel. The star–delta connection is not favoured of literature [19–21], two possible connections of star andparallel delta coils can be made: either thebecause two coils some practicalindifficulties in the of turns and the area ofconnection the conductors. are connected series or they arenumber connected in parallel. Thecross-section star–delta parallel is not Eventually, circulating occur, in resulting in additional losses. Therefore, area the series favoured because of somecurrents practicalmay difficulties the number of turns and the cross-section of the connection of the star and delta coils is adopted. The wiring connection of the series of star–delta conductors. Eventually, circulating currents may occur, resulting in additional losses. Therefore,coils the is shown in Figure 1. As equals 3,coils one is slot for theThe starwiring coil and two slots delta coils are series connection of the starq and delta adopted. connection of for the the series of star–delta coils is shown Figure 1. As q equals oneisslot for than the star slots by for athe delta are considered. Asin the current of the delta 3, coils lower the coil star and coil’stwo current factor ofcoils √3 , the considered. As the of the lower than the star coil’sofcurrent by turns a factor the number of turns ofcurrent the delta coilsdelta has coils to beishigher number star coil byofthe 3, same numberThis of turns of the delta coils has to be higher number of coils. star coil turns by the the same factor. factor. is to generate approximately the same than MMFthe with the two Consequently, number This is toofgenerate the same with thearea twoofcoils. Consequently, number of of turns the deltaapproximately coils is 45 turn/slot. TheMMF cross-section the delta coils mustthe be lower than turns of coil the delta coils Thegroups cross-section area of the coilsstar must becombined lower than the 3 . turn/slot. the star by a factor Two parallel are employed for delta both the and star– √ is 45 star coil by a factor 3. Two parallel groups are employed for both the star and combined star–delta delta windings. A single-layer winding is employed in both different windings. The phasor diagram windings. single-layer winding is employed inwindings both different windings. The 2. phasor of Nd are the of MMFs ofAboth the star and combined star–delta is sketched in Figure Ns anddiagram MMFs of both the star and combined star–delta windings is sketched in Figure 2. N and N are the s d number of turns of the star and delta coils, respectively. number of turns of the star and delta coils, respectively. Table 1. Synchronous Reluctance Motor (SynRM) parameters. Table 1. Synchronous Reluctance Motor (SynRM) parameters. Parameter Value Parameter Value Parameter Value Value Number of rotor flux barriers/pole 3 Active Parameter axial length 140 mm NumberNumber of rotor flux barriers/pole 336/2 Activesteel axial length 140 mm of stator slots/pole pairs Rotor M330-50A Number of stator slots/pole pairs 36/2 Rotor steel M330-50A Number of phases 3 Stator steel M270-50A Number of phases 3 Stator steel M270-50A Stator outer/inner 180/110 mm Rated 380 V380 V Stator outer/inner diameterdiameter 180/110 mm Rated voltage voltage Rotor shaft diameter mm Rated 5.5 kW Rotor shaft diameter 3535 mm Rated power power 5.5 kW Rotor outer diameter 109.4 mm Rated speed speed 3000 RPM Rotor outer diameter 109.4 mm Rated 3000 RPM Air gapAir length 0.30.3 mm Rated current current gap length mm Rated 12.23 12.23 A A

ic

c bc

ca a ab

ib

b

ia Figure 1. 1. Star–delta Star–delta coils coils connected connected in in series. series. Figure

The rotor of the SynRM is a transversally laminated type with three flux-barriers per pole as shown in Figure 3. The PM-assisted rotor is simply the SynRM rotor with inserted ferrite PMs in the centre of the flux-barrier as sketched in Figure 4. The ferrite PM type is Y30BH with a remanence (Br) and a coercive force (Hc) of 0.39 T and 234 kA/m, respectively. The steel grade of the machine’s core plays a major role in the losses and hence the efficiency of SynRMs, as proved in [7,8]. It is shown in [8] that the different steel materials result in different amounts of SynRM output power, and can increase the rated efficiency by 2.3% when using an NO20 grade instead of an M400-50A. However, the lower loss grades are more expensive both in raw material cost and in cutting cost [7]. In a rough approximation, the lowest loss grade will have more or less double the cost compared to the highest loss grade. In order to compromise between the

Energies 2017, 10, 1500

4 of 17

efficiency and the manufacturing cost of the prototypes, we have selected a lower loss grade for the stator core (M270-50A) than for the rotor core (M330-50A). This is because the stator core produces the majority of the iron losses of the SynRM. Energies 10,the 1500aforementioned details, four prototype SynRMs can be obtained. The abbreviations 4 of 174 of 18 From Energies 2017, 2017, 10, 1500 given in Table 2 are used in the remainder of the text.

efficiency and the manufacturing cost of the prototypes, we have selected a lower loss grade for the Energies 2017, 10, 1500 N Ithan for the rotor core (M330-50A). This is because the stator core produces 4 of 17 stator core (M270-50A) s c NdIca NsIc the majority of the iron losses of the SynRM. efficiency and the manufacturing cost of the prototypes, we have selected a lower loss grade for the From the aforementioned details, four prototype SynRMs can be obtained. The abbreviations stator core (M270-50A) than for the rotor core (M330-50A). This is because the stator core produces given in Table 2 are used in the remainder of the text. the majority of the iron losses of the SynRM. sIa prototype SynRMs can be obtained. NsIThe a From the aforementioned details,Nfour abbreviations N sIc 120° 30° N I given in Table 2 are used in the remainder of the text. d ca

NsIc

NdIbc

NsIc

NdIca NdIab

NsIc

NsIa

NsIb

NsIb

120° (a)

NsIa (b)

30°

NI

s a Nscombined Ia Figure 2. Phasor diagram of magnetomotive forces (MMFs) produced by star and star–

NdI(MMFs) Figure 2. Phasor diagram of120° magnetomotive forces produced by star and combined star–delta bc delta windings. (a) Star connection; (b) Combined star–delta connection. 30° NdIab windings. (a) Star connection; (b) Combined star–delta connection.

NdIbc NsIb NsIb NdIabflux-barriers per pole as The rotor of the SynRM is a transversally laminated type with three (a) (b) shown in Figure 3. The N PM-assisted rotor is simply the SynRM NsIb rotor with inserted ferrite PMs in the sIb Figure 2. Phasor diagram of(a)magnetomotive forces and combined star– centre of the flux-barrier as sketched in Figure 4. The(MMFs) ferrite produced PM(b)typeby is star Y30BH with a remanence (Br) delta windings. (a) Star connection; (b) Combined star–delta connection. and a coercive force (Hc) of 0.39 T and 234 kA/m, respectively. Figure 2. Phasor diagram of magnetomotive forces (MMFs) produced by star and combined star– delta windings. (a) Star connection; (b) Combined star–delta connection.

(a)

(b)

Figure 3. One pole geometry of S and Sd prototype SynRMs. (a) S prototype; (b) Sd prototype.

(a)

(b)

(a) (b) Figure 3. One pole geometry of S and Sd prototype SynRMs. (a) S prototype; (b) Sd prototype.

Figure 3. One pole geometry of S and Sd prototype SynRMs. (a) S prototype; (b) Sd prototype. Figure 3. One pole geometry of S and Sd prototype SynRMs. (a) S prototype; (b) Sd prototype. Figure 4. Flux-barriers with inserted ferrite permanent magnets (PMs). Table 2. SynRM abbreviations.

Machine Abbreviation Stator Winding Rotor Conventional star connection, Figure 3a Flux-barriers without PMs, Figure 3 S Combined star–delta connection, Figure 3b Flux-barriers without PMs, Figure 3 Sd Conventional star connection, Figure 3a inserted Flux-barrierspermanent with ferritemagnets PMs, Figure 4 S-PM Figure 4. 4. Flux-barriers (PMs). Figure Flux-barrierswith with inserted ferrite ferrite permanent magnets (PMs). Combined star–delta Figure 3binserted Flux-barriers ferrite PMs, Figure(PMs). 4 Sd-PM Figure 4. connection, Flux-barriers with ferritewith permanent magnets

Table Table2.2.SynRM SynRM abbreviations. abbreviations.

The steel grade of the machine’s core plays a major role in the losses and hence the efficiency Machine Machine Abbreviation Abbreviation of SynRMs, as proved in [7,8]. It is shown in [8] that the different steel materials result in different Stator Winding Rotor Stator Winding Rotor amounts of SynRM output power, and can increase the rated efficiency by 2.3% when an NO20 Conventional star connection,Figure Figure3a 3a Flux-barriers without Susing Conventional star connection, Flux-barriers withoutPMs, PMs,Figure Figure3 3 S grade instead of an M400-50A. However, the lower loss grades are more expensive both in raw material Combined star–delta connection,Figure Figure3b 3b Flux-barriers Flux-barriers without SdSd Combined star–delta connection, withoutPMs, PMs,Figure Figure3 3 Conventional star connection, Figure 3a Flux-barriers with ferrite PMs, Figure 4 S-PM Conventional connection, Flux-barriersthe with ferriteloss PMs,grade Figure will 4 S-PM cost and in cuttingstar cost [7]. In a Figure rough3a approximation, lowest have more or less Combined star–delta connection,Figure Figure3b 3b Flux-barriers Flux-barriers with PMs, Sd-PM star–delta connection, withferrite ferrite PMs,Figure Figure4between 4 Sd-PM doubleCombined the cost compared to the highest loss grade. In order to compromise the efficiency and the manufacturing cost of the prototypes, we have selected a lower loss grade for the stator core (M270-50A) than for the rotor core (M330-50A). This is because the stator core produces the majority of the iron losses of the SynRM. From the aforementioned details, four prototype SynRMs can be obtained. The abbreviations given in Table 2 are used in the remainder of the text.

Energies 2017, 10, 1500

5 of 18

Table 2. SynRM abbreviations. Machine Stator Winding Conventional star connection, Figure 3a Combined star–delta connection, Figure 3b Conventional star connection, Figure 3a Combined star–delta connection, Figure 3b

Abbreviation Rotor Flux-barriers without PMs, Figure 3 Flux-barriers without PMs, Figure 3 Flux-barriers with ferrite PMs, Figure 4 Flux-barriers with ferrite PMs, Figure 4

S Sd S-PM Sd-PM

3. Performance Comparison of SynRMs Using Finite Element Model (FEM) Energies 2017, 10, 1500

5 of 17

Four SynRMs are modelled using 2D-MAXWELL ANSYS software (16.2.0, ANSYS, INC., Berkeley, CA, USA). The goal is to compare their performance, i.e., output torque, torque ripple, power factor, Four SynRMsInare using software INC., losses, and efficiency. themodelled simulation, the2D-MAXWELL rotor is rotatedANSYS at a fixed speed.(16.2.0, In the ANSYS, stator, three-phase Berkeley, CA, USA). The goal is to compare their performance, i.e., output torque, torque ripple, sinusoidal currents are enforced into the windings to simply emulate the current-controlled inverter power factor, losses, and efficiency. In the simulation, the rotor is rotated at a fixed speed. In the that supplies the SynRM. For the Sd machines, the three sources are connected to the star coils as stator, three-phase sinusoidal currents are enforced into the windings to simply emulate the currentseen controlled Figure 1. inverter Consequently, the currents in For the the delta are not they computed that supplies the SynRM. Sd coils machines, the enforced; three sources areare connected to by the coupled FEMasand circuit of Figure 1. that delta triplen harmonics the star coils seen Figuremodel 1. Consequently, theNotice currents in in thethe delta coilscoils, are not enforced; they areof the current may occur. These circulating are taken into 1. account the computed by the coupled FEM andcurrents circuit model of Figure Notice in that insimulation. the delta coils, triplen harmonics of thethe current may occur. circulating are taken into account in the Figure 5 shows output torque of These the four SynRMscurrents as a function of the current angle at rated simulation. speed (3000 rpm) and for half- and full-rated current (12.23 A). For half-rated current, it is obvious Figure 5torque showsof thethe output torque of the four as aincreases function ofbythe current angle at rated and that the output Sd-PM, S-PM, and SdSynRMs machines about 41.85%, 34.55%, speed (3000 rpm) and for halfand full-rated current (12.23 A). For half-rated current, it is obvious 6.41%, respectively, compared to the S machine at the optimal current angles. The optimal current that the output torque of the Sd-PM, S-PM, and Sd machines increases by about 41.85%, 34.55%, and angle represents the angle of the stator’s current vector with respect to the d-axis that achieves the 6.41%, respectively, compared to the S machine at the optimal current angles. The optimal current maximum output torque. It is evident from Figure 5 that the optimal current angle is not a fixed angle represents the angle of the stator’s current vector with respect to the d-axis that achieves the valuemaximum and depends the stator’s current level and saturation the machine’s outputon torque. It is evident from Figure 5 thaton thethe optimal currentbehaviour angle is not of a fixed value core and as well. Thison can noticed in Figure 5 byon comparing the behaviour different curves of several machines depends thebe stator’s current level and the saturation of the machine’s core as and current levels. Furthermore, the output torque of the the different Sd-PM machine highermachines than the and S-PM by well. This can be noticed in Figure 5 by comparing curves ofisseveral the output of the Sd-PM machine is of higher than the in S-PM aboutcurrent 5.42%levels. at theFurthermore, optimal current angles.torque This means that the amount the increase theby output about 5.42% the optimal current angles. rotors This means that theand amount of the increase with in thePM-assisted output torque of the twoatmachines with reluctance (S and Sd) the two machines of and the two machines with reluctance rotors (S andisSd) the two This machines with PM-assisted rotorstorque (S-PM Sd-PM) at the optimal current angles notand constant. is because of the different rotors (S-PM and Sd-PM) at the optimal current angles is not constant. This is because of the different dq-axis currents and the saturation of the machine core. On the other hand, for full-rated current, it is dq-axis currents and the saturation of the machine core. On the other hand, for full-rated current, it is clear from Figure 5 that the output torque of the Sd, S-PM, and Sd-PM machines is higher than the S clear from Figure 5 that the output torque of the Sd, S-PM, and Sd-PM machines is higher than the S machine by about 5.02%, 17.01%, and 22.37%, respectively, at the optimal current angles. This can be machine by about 5.02%, 17.01%, and 22.37%, respectively, at the optimal current angles. This can be seen seen in Figure 6, in6,which thethe output thefour fourmachines machines is plotted for several positions. in Figure in which outputtorque torque of of the is plotted for several rotorrotor positions. An interesting observation here is that the increase in the output torque of the Sd, S-PM, and Sd-PM An interesting observation here is that the increase in the output torque of the Sd, S-PM, and Sd-PM machines compared to the S machine value;it itis is current-dependent. flux density machines compared to the S machineisisnot notaa constant constant value; current-dependent. The The flux density ◦ of Figure distribution of the four machines r =00° of Figure 66isisshown Figure 7. It7.isItclear that that the Sd-PM distribution of the four machines atatθ rθ= showninin Figure is clear the Sd-PM machine regions with much higherflux fluxdensity density compared other machines, in particular in machine has has regions with much higher comparedtotothe the other machines, in particular in the stator yoke. the stator yoke. 3. Performance Comparison of SynRMs Using Finite Element Model (FEM)

Figure 5. SynRM outputtorque torque(T (Tee)) as as aa function angle (α) (α) at rated speed. Figure 5. SynRM output functionofofcurrent current angle at rated speed.

Energies 2017, 10, 1500 Energies 2017, 10, 1500 Energies 2017, 10, 1500

6 of 18 6 of 17 6 of 17

Figure 6. SynRM SynRM output output torque torque (T (Te)) as function of of mechanical mechanical rotor angle (θ (θm)) at at rated rated conditions conditions Figure 6. as aa function rotor angle Figure 6. SynRM output torque (Tee) as a function of mechanical rotor angle (θmm) at rated conditions and optimal current angles. and optimal current angles. and optimal current angles.

(a) (a)

(b) (b)

(c) (c)

(d) (d)

Figure 7. Flux density distribution of the four prototypes at rated current and optimal current angles. Figure 7. Flux density distribution of the four prototypes at rated current and optimal current angles. Figure 7.Sd; Flux distribution (a) S; (b) (c)density S-PM; (d) Sd-PM. of the four prototypes at rated current and optimal current angles. (a) (a) S; S; (b) (b) Sd; Sd; (c) (c) S-PM; S-PM; (d) (d) Sd-PM. Sd-PM.

Figure 8 shows the output torque of the four machines as a function of the stator current at the Figure 88 shows the output torque of the machines as of stator current at Figure shows theand output torque of The the four four machines as aoutput a function function of the the current at the the optimal current angles rated speed. difference of the torque (instator percent) of the Sdoptimal current angles and rated speed. The difference of the output torque (in percent) of the Sdoptimal and rated speed. The of the output torquein(in percent) of the Sd-PM, PM, S-PM, andangles Sd machines compared to difference the S machine is reported Figure 9. Clearly, both PM, S-PM, andmachines Sd machines compared to the S machine is reported in9.Figure 9.both Clearly, both S-PM, and Sd compared to the S machine is reported in Figure Clearly, machines machines with PMs in the rotor (Sd-PM and S-PM) have much higher output torque compared to the machines in(Sd-PM the rotor (Sd-PM and S-PM) have much higher output torque compared to the with PMs with inThe thePMs rotor and S-PM) have much higher output compared to the S machine. S machine. Sd-PM machine has an increase in output torque torque of about 22.37% for rated current SThe machine. The Sd-PM machine has an increase in output torque of about 22.37% for rated current has ancurrent increasecompared in outputtotorque about 22.37% ratedthanks current about and Sd-PM of aboutmachine 150% for low the S of machine. This is for mainly toand the of inserted and offor about 150% forcompared low current compared to the S machine. This is mainly thanks toferrite the inserted 150% low current to the S machine. This is mainly thanks to the inserted PMs in ferrite PMs in the rotor and the improved winding factor of the star–delta connection. Furthermore, ferrite PMs inthe theimproved rotor and winding the improved winding factor ofconnection. the star–delta connection. Furthermore, the rotor and factor of the star–delta Furthermore, the difference in the difference in the output torque of the Sd-PM, S-PM, and Sd machines compared to the S machine the in theofoutput torque of theand Sd-PM, S-PM, and Sd machines to the S machine the difference outputwith torque the Sd-PM, S-PM, Sd machines thecompared S machine with decreases an increase in the stator current. This iscompared due to a to decrease in the decreases saliency factor decreases with an stator increase in theThis stator current. This is due to saliency a decrease in the saliencywith factor an increase in the current. is due to a decrease in the factor difference an difference with an increase in the current as shown in Figure 10. The saliency factor is the ratio difference with an increase in the current as shown in Figure 10. The saliency factor is the ratio increase in current asinductances shown in Figure The saliency factor the ratio the dlinkages and q axis between thethe d and q axis of the10. machine [22]. For lowiscurrent, thebetween dq-axis flux of between the dofand q machine axis inductances oflow the current, machinethe [22].dq-axis For low current, the of dq-axis flux linkages of inductances the [22]. For flux linkages the machine areratio low. the machine are low. Consequently, the machine torque component produced from the saliency the machine are low. Consequently, the machine torque from component produced from the low saliency ratio Consequently, machine component produced the from saliency is rather compared is rather low the compared totorque the torque component produced theratio PMs. This shows a huge is rather low compared to the torque component produced from the PMs. This shows a huge to the torque component produced from themachines PMs. Thiswith shows a huge difference between the output difference between the output torque of the PMs (Sd-PM and S-PM) compared to the difference between the with output torque of the machines with PMs (Sd-PM and S-PM) compared to the torque of the machines PMs (Sd-PM and S-PM) compared to the machines without PMs (S and machines without PMs (S and Sd). With an increasing stator current, the dq-axis flux linkages of Sd). the machines without PMs (Scurrent, and Sd). With an increasing stator current, theincrease, dq-axis flux linkages of the With an increasing stator the dq-axis flux linkages of the machine hence the difference machine increase, hence the difference in the torque between the PM machines and PM-free machines machine increase, hencethe thePM difference in the between the PM machines anddifference PM-free machines in the torque machines andtorque PM-free machines decreases. The in torque decreases. Thebetween difference in torque between the machines becomes almost constant at a high stator decreases. The difference in torque between the machines becomes almost constant at a high stator current because of the saturation of the material core. current because of the saturation of the material core.

Energies 2017, 10, 1500

7 of 18

between the machines becomes almost constant at a high stator current because of the saturation of the Energies 10, 77 of Energies 2017, 2017, 10, 1500 1500 of 17 17 material core. Energies 2017, 10, 1500

7 of 17

Figure 8. SynRM output torque e) as stator current (RMS) at current angles Figure of of stator current (RMS) at optimal current angles and Figure 8. 8. SynRM SynRMoutput outputtorque torque(T(T (T ) asa afunction a function function of stator current (RMS) at optimal optimal current angles e ) eas Figure 8. SynRM output torque (Te) as a function of stator current (RMS) at optimal current angles and rated speed. rated speed. and rated speed. and rated speed.

Figure 9. in torque % (T of stator current Figure 9. 9. Difference Difference in in the the output output torque torque % % (T (Tee)) as as aa function function of of stator stator current current (RMS) (RMS) with with respect respect to to Figure Figure 9. Difference Difference in the the output output torque % and (Tee)) as as aa function function ofthe stator current (RMS) (RMS) with with respect respect to to the S machine at the optimal current angles rated speed of SynRMs. the S machine at the optimal current angles and rated speed of the SynRMs. the the SS machine machine at at the the optimal optimal current current angles angles and and rated rated speed speed of ofthe theSynRMs. SynRMs.

Figure Figure 10. 10. Difference Difference in in the the saliency saliency factor factor % % (SR) (SR) as as aa function function of of stator stator current current (RMS) (RMS) with with respect respect Figure 10. Difference inoptimal thesaliency saliency factor %(SR) (SR) asa afunction function of stator current (RMS) with respect Figure 10. Difference in the factor % as of stator current (RMS) with respect to to the S machine at the current angles and rated speed of the SynRMs. to the S machine at the optimal current angles and rated speed of the SynRMs. to the S machine at the optimal current angles rated speed of the SynRMs. the S machine at the optimal current angles andand rated speed of the SynRMs.

Figure Figure 11 11 shows shows the the variation variation of of torque torque ripple ripple (in (in percent) percent) as as aa function function of of current current angle angle at at the the shows variation function current the Figure 11 shows the variation of torque ripple (in percent) as a function of current angle at the rated conditions of the four machines. It is observed that the torque ripple of the four machines rated conditions of the four machines. It is observed that the torque ripple of the four machines conditions the four four machines. machines. It is observed that the angle, the four four machines machines rated conditions of the torque ripple of the decreases decreases with with an an increase increase in in the the current current angle angle until until an an optimal optimal angle, and and then then increases increases again. again. The The withcurrent an the until anan optimal angle, andand thenthen increases again. The decreases an increase increaseinin thecurrent currentangle angle until optimal angle, increases again. value value and and the the current angle angle of of the the minimum minimum torque torque ripple ripple is is different different for for the the four four machines. machines. This This is is value and the current angle of theofminimum torque ripple ripple istodifferent forspatial thefor four machines. Thisthe is The value and thethat current angle the minimum torque is different the four machines. due to the fact the torque ripple is proportional both the harmonics of due to the fact that the torque ripple is proportional to both the spatial harmonics of the due is to due the tofact ripple is proportional to to both of the This thethat fact the that torque the torque ripple is proportional boththe thespatial spatial harmonics harmonics of the magnetomotive magnetomotive force force (MMF) (MMF) and and the the machines’ machines’ average average torque. torque. Both Both the the harmonics harmonics and and the the average average magnetomotive force (MMF) (MMF) and and the the machines’ machines’ average torque. torque. Both Both the harmonics harmonics and the average magnetomotive force average and torque torque of of the the four four machines machines are are different. different. By By comparing comparing the the torque torque ripple ripple of of the the four four machines, machines, it it can can the four machines are different. By comparing the torque ripple of the four machines, it can torque of the four machines are different. By comparing the torque ripple of the four machines, it be noticed that those machines with a star–delta connected stator have a higher torque ripple be noticed that those machines with a star–delta connected stator have a higher torque ripple be noticed that those machines with a star–delta connected stator have a higher ripple can be noticed that those machines with a star–delta connected stator have a higher torque ripple compared compared to to those those machines machines with with star star winding. winding. This This is is due due to to two two main main reasons. reasons. On On the the one one hand, hand, those machines This is On the compared to those machines with star winding. due to two main reasons. one hand, the first reason is the circulating current components of the delta coils. Although these components the first reason is the circulating current components of the delta coils. Although these components thenot first reason is to the circulating current components of the delta coils. Althoughripple’s these components do do not contribute contribute to average average torque torque production, production, they they negatively negatively affect affect the the torque torque ripple’s magnitude. magnitude. do not contribute to average torque production, they negatively affect the torque ripple’s magnitude. The The currents currents in in the the star star and and delta delta coils coils of of the the different different connections connections at at the the rated rated conditions conditions and and optimal optimal The currents inare thereported star andin delta coils12. ofThe the different connections at theas rated conditions and optimal current angle Figure star currents are enforced pure sinusoidal currents— current angle are reported in Figure 12. The star currents are enforced as pure sinusoidal currents— current angle are reported in Figure 12. The star currents are enforced as pure sinusoidal currents— as as mentioned mentioned before—in before—in all all the the different different connections, connections, while while the the delta delta currents currents in in the the combined combined star– star– as mentioned before—in all the different connections, while the delta currents in the combined star–

Energies 2017, 10, 1500 Energies 2017, 10, 1500

8 of 17 8 of 17

delta delta windings windings (Sd (Sd and and Sd-PM) Sd-PM) are are computed computed based based on on FEM. FEM. ItIt is is evident evident that that the the delta delta coils coils have have circulating circulating current current components. components. The The harmonic harmonic spectrum spectrum of of the the currents currents is is reported reported in in Figure Figure 13. 13. Energies 2017, 10, 1500 8 of 18 Apart Apart from from the the fundamental fundamental component, component, the the dominant dominant harmonic harmonic component component is is the the 3rd 3rd in in both both Sd Sd and Sd-PM, and its value is approximately the same: about 11.2% of the fundamental component and Sd-PM, and its value is approximately the same: about 11.2% of the fundamental component of of the connection. These harmonics negatively the ripple, as first reason is the circulating current components of theaffect delta coils. Although these components the Sd Sd machine’s machine’s connection. These harmonics negatively affect the torque torque ripple, as observed observed in in Figure the of ripple % Sd is do not 11. contribute to average torque production, they negatively affectthe theSd-PM torqueand ripple’s magnitude. Figure 11. Notice Notice that that the difference difference of the the torque torque ripple % between between the Sd-PM and Sd machines machines is due to the difference in the harmonic components as seen in Figure 13. On the other hand, the second The in the star delta coilscomponents of the different connections theOn rated due currents to the difference in and the harmonic as seen in Figureat13. the conditions other hand,and theoptimal second reason due fact angles have been selected on currentis are reported in the Figure 12.flux-barrier The star currents enforced as pure sinusoidal currents—as reason isangle due to to the the fact that that the rotor rotor flux-barrier anglesare have been optimally optimally selected based based on the the machines with a star connection. The influence of these angles is mainly on the torque ripple of the mentioned before—in all the different connections, while the delta currents in the combined star–delta machines with a star connection. The influence of these angles is mainly on torque ripple of the SynRMs shown in ititis to the SynRM rotor flux-barrier angles windings (Sd and Sd-PM) are means computed on FEM. is evident that the delta coils have circulating SynRMsas as shown in[6]. [6].This This meansthat thatbased ispossible possible toItoptimize optimize the SynRM rotor flux-barrier angles with respect to the combined star–delta connection, and hence the torque ripple can be decreased. current components. The harmonic spectrum of the currents is reported in Figure from the with respect to the combined star–delta connection, and hence the torque ripple 13. canApart be decreased. The torque in 11 about 6.44% 9.5% (star–delta fundamental component, dominant harmonic is the 3rd in to both Sd and and The torque ripple ripple in Figure Figurethe 11 increases increases from from aboutcomponent 6.44% (star (star connection) connection) to about about 9.5%Sd-PM, (star–delta connection). its value is approximately the same: about 11.2% of the fundamental component of the Sd machine’s connection). For Figure 14 the variation the ripple percent) different connection. These harmonics negatively affect torqueof ripple, as observed in 11.for Notice that Forthe thefour fourmachines, machines, Figure 14shows shows thethe variation of thetorque torque ripple(in (inFigure percent) for different stator currents at rated conditions and optimal current angles. It is seen that the SynRMs’ torque the difference torque ripple %and between thecurrent Sd-PM and Sd machines duethe to SynRMs’ the difference in stator currentsofatthe rated conditions optimal angles. It is seen is that torque ripple decreases with stator current. of the in the average the harmonic components as seen in Figure 13.This On is the other because hand, the reason due to the ripple decreases with increasing increasing stator current. This is mainly mainly because ofsecond the increase increase inis the average torque with current and fact that the is as percentage of average fact that the the rotor flux-barrier have been selected based the machines with a star torque with the stator stator currentangles and the the fact thatoptimally the ripple ripple is given given as aaon percentage of the the average torque. In an absolute peak-to-peak value, the ripple increases linearly with an increase in the stator connection. influence of these angles is mainly on the torque ripple the SynRMs torque. In anThe absolute peak-to-peak value, the ripple increases linearly withofan increase in as theshown stator current, as However, the aa smaller effect on torque in [6]. This means thatin it [22]. is possible to optimize the SynRMvalue rotorhas flux-barrier respect to current, as presented presented in [22]. However, the peak-to-peak peak-to-peak value has smallerangles effectwith on the the torque ripple than the average torque. In addition, the Sd machines have a higher torque ripple than the the combined connection, and hence ripple cana be decreased. torque ripple than thestar–delta average torque. In addition, thethe Sd torque machines have higher torque The ripple thanripple the SS machines the connection as before. in Figure with 11 increases about 6.44% (star connection) machines with the star starfrom connection as explained explained before. to about 9.5% (star–delta connection).

Figure 11. SynRM ripple (Tr %) at rated current and speed. SynRMtorque %)as asaaafunction functionof ofcurrent currentangle angle(α) Figure 11. SynRM torque ripple (Trr %) as function of current angle (α) at rated current and speed.

Figure 12. Currents of star and delta coils at rated conditions and optimal current angles. Figure angles. Figure 12. 12. Currents Currents of of star star and and delta delta coils coils at at rated rated conditions conditions and and optimal optimal current current angles.

Energies 2017, 10, 1500 Energies 2017, 10, 1500

9 of 18 9 of 17

Energies 2017, 10, 1500

9 of 17

Figure 13. Harmonic rated conditions conditions and optimal Figure 13. Harmonic spectrum spectrum of of currents currents in in star star and and delta delta coils coils at at rated and optimal current angles. current angles.

For the four machines, Figure 14 shows the variation of the torque ripple (in percent) for different stator currents at rated conditions and optimal current angles. It is seen that the SynRMs’ torque ripple decreases with increasing stator current. This is mainly because of the increase in the average torque with the stator current and the fact that the ripple is given as a percentage of the average torque. In an bsolute peak-to-peak value, the ripple increases linearly with an increase in the stator current, as presented in [22]. However, the peak-to-peak value has a smaller effect on the torque ripple than the Figure 13. Harmonic spectrum currents have in star and delta coils ripple at rated conditions and optimal average torque. In addition, the Sdof machines a higher torque than the S machines with the current angles. star connection as explained before.

Figure 14. SynRM torque ripple % (Tr) as a function of stator current (RMS) at optimal current angles and rated speed.

The power factor of the four SynRMs as a function of the current angles for rated conditions is shown in Figure 15. It is observed that the SynRMs’ power factor increases with an increase in the current angle until an optimal value. This is because of the increase in the saliency ratio. Note that the maximum value is at a higher current angle than for the maximum torque in Figure 5. Figure 15 confirms findings in other studies in the literature, e.g., [16], that adding PMs in the rotor increases the power factor dramatically. However, the figure shows that there is almost no influence on a machine’s factor when using star–delta connection of current the conventional Figurepower 14. SynRM torque ripple % (T (Tar))combined as a function of stator current (RMS)instead at optimal angles Figure 14. SynRM torque ripple % r as a function of stator current (RMS) at optimal current angles and rated speed. star connection, both for the machines with and without ferrite PMs. This is because the combined and rated speed. star–delta winding has a non-significant influence on the phase shift between the stator current and power factor rated conditions is Thevectors. factor of of the the four fourSynRMs SynRMsas asaafunction functionofofthe thecurrent currentangles anglesforfor rated conditions voltage shown in in Figure 15.15. It isIt observed thatthat thethe SynRMs’ power factor increases withwith an increase in the is shown Figure is observed SynRMs’ power factor increases an increase in current angle untiluntil an optimal value. ThisThis is because of the in the ratio.ratio. NoteNote that the current angle an optimal value. is because of increase the increase in saliency the saliency the maximum valuevalue is at a is higher currentcurrent angle than forthan the maximum torque in torque Figure in 5. Figure that the maximum at a higher angle for the maximum Figure15 5. confirms findings in other studies in the literature, e.g., [16], that adding PMs in the rotor increases Figure 15 confirms findings in other studies in the literature, e.g., [16], that adding PMs in the rotor the powerthe factor dramatically. However, the figure that there almost no influence on a increases power factor dramatically. However, theshows figure shows thatisthere is almost no influence machine’s power factor when using a combined star–delta on a machine’s power factor when using a combined star–deltaconnection connectioninstead insteadof ofthe the conventional star connection, both for the machines with and without without ferrite PMs. This is because the combined star–delta winding has a non-significant influence on the phase shift between the stator current and voltage vectors.

Figure 15. SynRM power factor (PF) as a function of current angle (α) at rated current and speed.

The variation of the power factor of the four SynRMs as a function of the stator current is reported in Figure 16. The simulations are done at the optimal current angles and rated speed. Notice

the maximum value is at a higher current angle than for the maximum torque in Figure 5. Figure 15 confirms findings in other studies in the literature, e.g., [16], that adding PMs in the rotor increases the power factor dramatically. However, the figure shows that there is almost no influence on a machine’s power factor when using a combined star–delta connection instead of the conventional star connection, both for the machines with and without ferrite PMs. This is because the combined star–delta winding has a non-significant influence on the phase shift between the stator current and 10 of 18 Energies 2017, 10, 1500 voltage vectors.

Figure 15. SynRM power factor (PF) as a function of current angle (α) at rated current and speed.

Figure 15. SynRM power factor (PF) as a function of current angle (α) at rated current and speed. The variation of the power factor of the four SynRMs as a function of the stator current is

Thereported variation of the16. power factor of the four at SynRMs as acurrent function of the current is reported in Figure The simulations are done the optimal angles andstator rated speed. Notice Energies 2017, 10, 1500 10 of 17 in Figure 16. The simulations are done at the optimal current angles and rated speed. Notice Energies 2017, 10, 1500 10 of 17 that the step variation thevariation power factor in Figure 16inisFigure due to variation of the optimal current angle based that thein step in the power factor 16the is due to the variation of the optimal current that the step variation in the power factor in Figure 16from is due to the variation of the optimal current on the stator’s current Wemagnitude. already know Figure 15 Figure that the ferrite the angle based on themagnitude. stator’s current We already know from 15 that the PMs ferriteincrease PMs angle based on the stator’s current magnitude. We already know from Figure 15 that the ferrite PMs increase the power factor significantly. addition, learn 16 from Figure 16 that gain in power power factor significantly. In addition, weIn learn fromwe Figure that the gain inthe power factor becomes increase the power factor significantly. In addition, we learn from Figure 16 that the gain in power factor becomes lower at high stator currents. This is because ofthe the increase in the flux linkage of the lower atfactor highbecomes stator currents. Thisstator is because of This the increase linkage thelinkage machine, resulting lower at high currents. is becausein of the flux increase in theof flux of the machine, resulting in an increase in the phase angle between the voltage and current vectors. This in an increase the phase between the voltage and current vectors. This can be understood machine,inresulting in anangle increase in the phase angle between the voltage and current vectors. This can be understood simply from the machine vector diagram shown in Figure 17. The voltage and be the understood simply from the machine vector diagram17. shown Figure 17. The voltageequations and simply can from machine vector diagram shown in Figure Theinvoltage and torque of torque equations of PMaSynRM are be given by: torque equations of PMaSynRM are be given by: PMaSynRM are be given by: ( ωeψψq + V = R I ++ω pm V ω +ωωeψ eψ Vd dd== RRsssIIddd+ eeψ qq + ωeψ pmpm (1) (1) (1) R I +ω ψ  VVq q== R V = R sIIq ++ωωeψe ψdd



q

s q

e

d

3 =3 P P((ψ − ψψ II + + ψψpm (2) ψ II − (2) TTee = pmIIdd )) Te = 2 P (ψ ddd Iqqq − ψ qqI ddd + ψ pm Id ) (2) 2 where ψwhere represents the flux-linkage, I isI isthe stator voltage, Rs stator is the stator ψ represents the flux-linkage, thestator stator current, current, VVis is thethe stator voltage, Rs is the where ψ represents the flux-linkage, I is the stator current, V is the stator voltage, Rs is the stator resistance ω e is the electrical speed of the PMaSynRM. The subscript symbols d, q, and pm are pm are resistance and ωand is the electrical speed of the PMaSynRM. The subscript symbols d, q, and e resistance and ωe is the electrical speed of the PMaSynRM. The subscript symbols d, q, and pm are direct, quadrature and permanent magnet, respectively. direct, quadrature and permanent magnet, direct, quadrature and permanent magnet,respectively. respectively.

Figure 16. SynRM power factor (PF) as a function of stator current (RMS) at optimal current angles

Figure 16. SynRM power factor (PF) function of (RMS) at optimal currentcurrent angles angles Figure 16. SynRM power factor (PF) asasa afunction ofstator statorcurrent current (RMS) at optimal and rated speed. andspeed. rated speed. and rated

q-axis q-axis RsId ωeψpm -ωeψqRsId ωeψpm -ωeψq ωψ ωeeψdd I Iqq RI RssIqq

V Vmm

ϕ ϕ α α

I Imm I Idd

d-axis d-axis

Figure 17. Vector diagram of PMaSynRM. Figure 17. Vector diagram of PMaSynRM.

Figure 17. Vector diagram of PMaSynRM. The simulated efficiency of the four SynRMs as a function of the stator current at the optimal The simulated efficiency of the four SynRMs as a function of the stator current at the optimal current angles and for half- and full-rated speed is reported in Figure 18. The efficiency calculation current angles and for half- and full-rated speed is reported in Figure 18. The efficiency calculation includes only the copper and iron losses of the machine. The copper losses are obtained using the includes only the copper and iron losses of the machine. The copper losses are obtained using the measured winding resistance of the machine and the current magnitude, which are equal for star and measured winding resistance of the machine and the current magnitude, which are equal for star and combined star-delta connections. This is because the increase in the number of turns of the delta coils combined star-delta connections. This is because the increase in the number of turns of the delta coils is compensated for by a reduction of same factor ( 3 ) in the cross-section area. The iron losses are is compensated for by a reduction of same factor ( 3 ) in the cross-section area. The iron losses are computed using the magnetic flux density B resulting from the FEM calculations for several points computed using the magnetic flux density B resulting from the FEM calculations for several points

Energies 2017, 10, 1500

11 of 18

The simulated efficiency of the four SynRMs as a function of the stator current at the optimal current angles and for half- and full-rated speed is reported in Figure 18. The efficiency calculation includes only the copper and iron losses of the machine. The copper losses are obtained using the measured winding resistance of the machine and the current magnitude, which are equal for star and combined star-delta connections. This is because the√increase in the number of turns of the delta coils is compensated for by a reduction of same factor ( 3) in the cross-section area. The iron losses are computed using the magnetic flux density B resulting from the FEM calculations for several points and positions. Then, the iron losses are obtained as in [8]. Figure 18 shows a slight increase in the Energies 2017, 10, of machines’ points Energies 2017,efficiency 10, 1500 1500 using the combined star-delta winding instead of the star one: about 0.33% 11 11 of 17 17 higher at maximum. Moreover, the efficiency of the machine is increased slightly when inserting PMs inserting PMs rotor. This is efficiency the with into the rotor. Thisthe is clear thecomparing efficiency the of the Sd-PMof machine withmachine that of the S inserting PMs into into the rotor.when This comparing is clear clear when when comparing the efficiency of the Sd-PM Sd-PM machine with that of the S machine: about 1.25% points higher for half-rated speed and 0.82% points for full-rated machine: 1.25% points for half-rated and 0.82% points full-rated speed. The low that of theabout S machine: about higher 1.25% points higher speed for half-rated speed andfor 0.82% points for full-rated speed. The the between can be from 19, difference thedifference efficiency in between the machines canthe be machines understood from Figure 19, which shows speed. Theinlow low difference in the efficiency efficiency between the machines can be understood understood from Figure Figure the 19, which shows the computed total losses of the four machines for halfand full-rated speeds. The computed totalthe losses of the four full-rated The strong increase which shows computed totalmachines losses of for thehalffourand machines for speeds. half- and full-rated speeds.with The strong indicates the copper losses are for currentincrease indicateswith that current the copper losses that (which the same for(which the machines) are dominant. It is clear strong increase with current indicates that theare copper losses (which are the the same same for the the machines) machines) are dominant. It is clear that the losses are approximately similar; only a slight increase in the losses thatdominant. the losses It are similar; a slight increase the losses the SynRMs are is approximately clear that the losses are only approximately similar;inonly a slightofincrease in thehaving losses of the SynRMs having combined star–delta windings occurs due to circulating harmonic currents. combined star–delta windings occurs due towindings circulating harmonic of the SynRMs having combined star–delta occurs due tocurrents. circulating harmonic currents.

Figure 18. The stator current current (RMS) (RMS) at at optimal optimal current current angles angles simulated efficiency function of Figure 18. The simulated simulated efficiency as as aa function of stator (only copper copper and and iron iron losses losses are are taken into account). (only are taken taken into into account). account).

Figure 19. The simulated losses (copper as aa function of stator current (RMS) optimal Figure 19. The The simulated simulated total total losses losses (copper (copper ++ + iron) iron) (RMS) at at optimal optimal Figure 19. total iron) as as a function function of of stator stator current current (RMS) at current angles. current angles. current angles.

4. 4. Experimental Experimental Validation Validation For For the the validity validity of of the the simulated simulated results results presented presented above, above, two two different different stators stators and and rotors rotors are are manufactured. The two stators have similar geometrical parameters: one has conventional star manufactured. The two stators have similar geometrical parameters: one has conventional star winding winding and and the the second second contains contains the the star–delta star–delta winding. winding. The The two two rotors rotors are are flux-barrier flux-barrier type type and and have four poles with three flux-barriers per pole. The rotors have similar geometrical parameters: have four poles with three flux-barriers per pole. The rotors have similar geometrical parameters: one one

Energies 2017, 10, 1500

12 of 18

4. Experimental Validation For the validity of the simulated results presented above, two different stators and rotors are manufactured. The two stators have similar geometrical parameters: one has conventional star winding and the second contains the star–delta winding. The two rotors are flux-barrier type and have four poles with three flux-barriers per pole. The rotors have similar geometrical parameters: one rotor with ferrite PMs inserted in the center of the flux-barriers (Figure 4), and a second one without PMs. Four prototype SynRMs can be obtained using the two stators and the two rotors with the parameters given in Table 1. Photographs of the prototypes and the complete experimental setup are reported in Figures Energies 2017,20 10,and 150021, respectively. 12 of 17

Energies 2017, 10, 1500

12 of 17

Sd-stator Sd-stator

Rel-rotor Rel-rotor

One rotor One rotor lamination lamination

Rel-PM rotor Rel-PM rotor

S-stator S-stator Ferrite-PM Ferrite-PM

Figure 20. A photograph of the prototypes, where S is a conventional star-connected stator, Sd is a Figure 20. 20. AAphotograph photographofofthe theprototypes, prototypes,where whereS Sisisa aconventional conventionalstar-connected star-connected stator, Figure stator, SdSd is is a combined star-delta-connected stator, Rel isisaaconventional rotor without PMs, and Rel-PM is aa rotor a combined star-delta-connected stator, Rel conventional rotor without PMs, and Rel-PM is rotor combined star-delta-connected stator, Rel is a conventional rotor without PMs, and Rel-PM is a rotor with PMs. PMs. with with PMs. Induction Induction motor motor PC PC

Torque sensor Torque sensor SynRM SynRM

Power Power analyzer analyzer Inverter Inverter Ds1103 Ds1103

DC supply DC supply

Figure 21. A photograph of the complete experimental setup. PC: personal computer; DC: direct Figure 21. A photograph of the complete experimental setup. PC: personal computer; DC: direct current. Figure 21. A photograph of the complete experimental setup. PC: personal computer; DC: current. direct current.

In the experimental setup, a 9.3 kW induction motor is controlled by a three-phase industrial In the experimental setup, a 9.3 kW induction motor is controlled by a three-phase industrial inverter and employed tosetup, driveathe at the desired The SynRM is driven industrial in torque In the experimental 9.3 SynRM kW induction motor isspeed. controlled by a three-phase inverter and employed to drive the SynRM at the desired speed. The SynRM is driven in torque control by controlling a conventional voltageThe source inverter using pulse control widthinvertermode and employed to drive the SynRM atthree-phase the desired speed. SynRM is driven inatorque control mode by controlling a conventional three-phase voltage source inverter using a pulse widthmodulated technique. The direct current (DC) bus voltage andinverter the switching of the inverter mode by controlling a conventional three-phase voltage source using afrequency pulse width-modulated modulated technique. The direct current (DC) bus voltage and the switching frequency of the inverter are fixed atThe 600direct V andcurrent 6.6 kHz, respectively, allthe of switching the measurements. dSpace 1103 technique. (DC) bus voltagefor and frequencyThe of the inverter areplatform fixed at are fixed at 600 V and 6.6 kHz, respectively, for all of the measurements. The dSpace 1103 platform is employed to control the inverter of the SynRM. To measure the output torque of the SynRM, a is employed to control the inverter of the SynRM. To measure the output torque of the SynRM, a torque sensor is mounted between the induction motor and the SynRM. The electric input power is torque sensor is mounted between the induction motor and the SynRM. The electric input power is measured using a power analyser. An incremental encoder is used to measure the motor speed. measured using a power analyser. An incremental encoder is used to measure the motor speed. Figure 22 shows the measured and simulated output torque of the four prototypes as a function Figure 22 shows the measured and simulated output torque of the four prototypes as a function of the current angle at half-rated current and speed. The simulated and measured results correspond of the current angle at half-rated current and speed. The simulated and measured results correspond

Energies 2017, 10, 1500

13 of 18

600 V and 6.6 kHz, respectively, for all of the measurements. The dSpace 1103 platform is employed to control the inverter of the SynRM. To measure the output torque of the SynRM, a torque sensor is mounted between the induction motor and the SynRM. The electric input power is measured using a power analyser. An incremental encoder is used to measure the motor speed. Figure 22 shows the measured and simulated output torque of the four prototypes as a function of the current angle at half-rated current and speed. The simulated and measured results correspond very well. The measured and simulated power factor and output torque of the four SynRMs as a function of the stator current at the optimal current angles and rated speed are reported in Figures 23 and 24, respectively. Good matching between the simulated and measured results is noticed. It is noticed that Energies 2017, 10, 1500 13 of 17 the discrepancy between the simulated and measured results is about 5% at maximum. Energies 2017, 10, 1500 13 of 17

Figure 22. The The outputtorque torque (T)e)ofofthe thefour four SynRMs a function of the current angle at halfFigure SynRMs as as aasfunction of the angle (α) at(α) half-rated e e) of the Figure 22. 22. Theoutput output torque(T(T four SynRMs a function of current the current angle (α) at halfrated current and speed. current and speed. rated current and speed.

Figure 23. SynRM power factor (PF) as a function of stator current (RMS) at the optimal current angles Figure 23. 23. SynRM SynRM power power factor factor (PF) (PF) as as aa function of stator current (RMS) at the optimal current angles Figure and rated speed. (a) Sd-PM; (b) S-PM; (c)function Sd; (d) S.of stator current (RMS) at the optimal current angles and rated speed. (a) Sd-PM; (b) S-PM; (c) Sd; (d) S. and rated speed. (a) Sd-PM; (b) S-PM; (c) Sd; (d) S.

Figure 23. SynRM power factor (PF) as a function of stator current (RMS) at the optimal current angles Energies 2017, 10, speed. 1500 (a) Sd-PM; (b) S-PM; (c) Sd; (d) S. and rated

14 of 18

Figure Figure 24. 24. SynRM SynRM output output torque torque (T (Tee))as asaafunction functionof ofstator statorcurrent current(RMS) (RMS) at atthe theoptimal optimal current current angles angles and rated speed. (a) Sd-PM; (b) S-PM; (c) Sd; (d) S. and rated speed. (a) Sd-PM; (b) S-PM; (c) Sd; (d) S.

The measured total losses losses of the four prototypes as a function of the stator current at full-rated speed is shown shown in inFigure Figure25. 25.The Themeasured measured losses difference between measured output losses areare thethe difference between the the measured output and and input powers of the machines. The difference in losses of the four prototypes is not significant, input powers of the machines. The difference in losses of the four prototypes is not significant, similar Energies 2017, 10, 1500 14 of 17 similar to theoftrends of the simulated results. However, the simulated than the to the trends the simulated results. However, the simulated losses arelosses lower are thanlower the measured losses. Thislosses. is because, simulation, thesimulation, mechanicalthe andmechanical pulse widthand modulation (PWM) losses are measured Thisinisthe because, in the pulse width modulation not considered. In addition, the computed iron losses may be underestimated because the degradation (PWM) losses are not considered. In addition, the computed iron losses may be underestimated of the material propertiesofbythe cutting andproperties press fitting not included. because the degradation material byiscutting and press fitting is not included.

Figure 25. losses of the fourfour prototypes at optimal current anglesangles and rated Figure 25. The Themeasured measured losses of the prototypes at optimal current and speed rated (3000 rpm). speed (3000 rpm).

26 reports reportsthe themeasured measuredefficiency efficiencyofof the four prototypes several loading currents at Figure 26 the four prototypes forfor several loading currents at the the optimal current angles and the speed rated speed (3000 Itrpm). It is clear the efficiency of the optimal current angles and at theatrated (3000 rpm). is clear that the that efficiency of the SynRMs SynRMs improves slightly the star–delta winding andsignificantly improves significantly by adding PMs improves slightly using the using star–delta winding and improves by adding PMs in the rotor. in the rotor.machine The Sd-PM hasefficiency: the highestabout efficiency: at the rated This is The Sd-PM hasmachine the highest 93.60%about at the93.60% rated current. Thiscurrent. is higher than higher than the requiredfor minimum for the IE4 super premium efficiency class92.50% [23]: about for the required minimum the IE4 super premium efficiency class [23]: about for a92.50% four-pole a four-pole 5.5 kWmotor. induction The rated for efficiency formachines the other is: machines for the S 5.5 kW induction Themotor. rated efficiency the other 92.10%is: for92.10% the S machine, machine, the Sd and machine, and for machine. the S-PM machine. 92.36% for92.36% the Sdfor machine, 93.29% for93.29% the S-PM

Figure 26 reports the measured efficiency of the four prototypes for several loading currents at the optimal current angles and at the rated speed (3000 rpm). It is clear that the efficiency of the SynRMs improves slightly using the star–delta winding and improves significantly by adding PMs in the rotor. The Sd-PM machine has the highest efficiency: about 93.60% at the rated current. This is higher than the required minimum for the IE4 super premium efficiency class [23]: about 92.50% for four-pole Energies 2017,a 10, 1500 5.5 kW induction motor. The rated efficiency for the other machines is: 92.10% for the S machine, 92.36% for the Sd machine, and 93.29% for the S-PM machine.

15 of 18

Figure 26. The measured efficiency of the four prototypes at optimal current angles and rated speed

Figure 26. The measured efficiency of the four prototypes at optimal current angles and rated speed (3000 rpm). (3000 rpm). Figures 27–30 report the measured efficiency maps of the whole drive system (prototype +

Figures 27–30atreport measured efficiency maps of the drive system (prototype inverter) optimalthe current angles for speeds and currents up towhole the rated values (3000 rpm, 12.23 A). + inverter) A shown before, the maximum output torque of the four machines is different and the Sd-PM at optimal current angles for speeds and currents up to the rated values (3000 rpm, 12.23 A). A shown machine gives the highest output torque. In general, and in correspondence with the literature [15], before, the maximum output torque of the four machines is different and the Sd-PM machine gives adding ferrite PMs in the rotor of a SynRM increases the machine’s efficiency. It is worth noting that the highest torque. In general, in 30) correspondence with theother literature [15], ferrite theoutput efficiency of the Sd-PM machineand (Figure is much better than for the machines overadding the PMs in thewhole rotoroperating of a SynRM the atmachine’s It isthe worth the efficiency of range,increases but especially low loads. efficiency. This is because outputnoting torque that of Sd-PM is higher (Figure than the other machines the same happens especially for the low currents the Sd-PMmuch machine 30) is much for better thancurrents. for theThis other machines over whole operating as depicted in Figure 10. By comparing the machines regarding the winding configuration, the range, but especially at low loads. This is because the output torque of Sd-PM is much higher than the machines with combined star–delta windings have better efficiency compared to the machines with other machines for the same happens forThis lowiscurrents as the depicted in Figure 10. the conventional star currents. connection,This especially underespecially partial loads. because of increased By comparing the machines regarding the winding configuration, the machines with combined star–delta torque-to-current ratio. windings have better efficiency compared to the machines with the conventional star connection, especially under partial loads. This is because of the increased torque-to-current ratio. 15 of 17 Energies 2017, 10, 1500 Energies 2017, 10, 1500

15 of 17

Figure 27. Measured efficiency map of the whole drive system using the S machine at optimal current angles.

Figure 27. Measured efficiency map of the whole drive system using the S machine at optimal 27. Measured efficiency map of the whole drive system using the S machine at optimal current angles. currentFigure angles.

Figure 28. Measured efficiency map of the whole drive system using the Sd machine at optimal current angles. Figure 28. Measured efficiency map of the whole drive system using the Sd machine at optimal 28. Measured efficiency map of the whole drive system using the Sd machine at current angles.

Figure current angles.

optimal

Figure 28. Measured efficiency map of the whole drive system using the Sd machine at optimal 16 of 18 current angles.

Energies 2017, 10, 1500

29. Measured efficiency map of the whole whole drive drive system system using the the S-PM S-PM machine machine at at optimal optimal Figure 29. Energies 2017, 10, 1500 current angles. current angles.

16 of 17

Figure 30. 30. Measured Measured efficiency efficiency map map of of the thewhole whole drive drive system system using using the the Sd-PM Sd-PM machine machine at at optimal optimal Figure current angles. angles. current

5. Conclusions Conclusions 5. Thispaper paperinvestigates investigates and compares performance (output torque, This and compares the the performance (output torque, powerpower factor, factor, torque torque ripple, ripple, and efficiency) of four prototype SynRMs with identical geometry of the stator and and efficiency) of four prototype SynRMs with identical geometry of the stator and rotor stacks; therotor two stacks; the two stators have different winding layouts and the two rotors differ in having ferrite PMs. stators have different winding layouts and the two rotors differ in having ferrite PMs. The winding The winding are the starthe connection the combined winding, layouts are the layouts conventional starconventional connection and combinedand star–delta winding, star–delta while the rotors are the rotors aretype a flux-barrier rotor type withPMs. and without ferrite PMs. awhile flux-barrier rotor with and without ferrite For the the same same copper copper volume volume and and current, current, the the machine machine with with star–delta star–delta winding winding and and ferrite ferrite PMs PMs For inserted into the rotor (Sd-PM) corresponds to an approximately 22% increase in torque at rated inserted into the rotor (Sd-PM) corresponds to an approximately 22% increase in torque at rated currentand andspeed speedcompared compared machine with a conventional connection a reluctance current toto thethe machine with a conventional star star connection and aand reluctance rotor. rotor. This enhancement is mainly thanks to adding the ferrite PMs in the rotor and the improvement This enhancement is mainly thanks to adding the ferrite PMs in the rotor and the improvement in the in the winding of the combined winding. the Moreover, the is torque gain is currentwinding factor offactor the combined star–delta star–delta winding. Moreover, torque gain current-dependent: it dependent: it increases up to 150% for low current compared to the conventional star connection with increases up to 150% for low current compared to the conventional star connection with a reluctance a reluctance rotor. An interesting here is that themachine efficiency machine with a rotor. An interesting observation hereobservation is that the efficiency of the withofa the combined star–delta combined star–delta connection and(Sd-PM) a PM-assisted (Sd-PM) is high for partialitloads. In addition, connection and a PM-assisted rotor is highrotor for partial loads. In addition, can reach 93.60% it can reach 93.60% for rated conditions, which is higher than the required minimum for the IE4 super for rated conditions, which is higher than the required minimum for the IE4 super premium efficiency premium efficiency class. This is because of the increased torque-to-current ratio. Consequently, this class. This is because of the increased torque-to-current ratio. Consequently, this machine (Sd-PM) machine is acandidate very promising candidate for several industriale.g., applications, e.g., systems PV pumping is a very (Sd-PM) promising for several industrial applications, PV pumping and systems and automotive applications. automotive applications. On the other hand, there is a non-significant influence on the power factor and losses of SynRMs using different winding connections. The theoretical findings are experimentally validated using four identical prototype machines with different rotors and winding layouts. Acknowledgments: The authors acknowledge the Egyptian Ministry of Higher Education (Cultural Affairs and

Energies 2017, 10, 1500

17 of 18

On the other hand, there is a non-significant influence on the power factor and losses of SynRMs using different winding connections. The theoretical findings are experimentally validated using four identical prototype machines with different rotors and winding layouts. Acknowledgments: The authors acknowledge the Egyptian Ministry of Higher Education (Cultural Affairs and Missions Sector) and Special Research Fund of Ghent University (BOF) for the financial support during this work. Author Contributions: All the authors contributed substantially to the work presented. Mohamed Nabil Fathy Ibrahim did the simulation and experimental works. In addition, he wrote the paper. Peter Sergeant gave a conceptual approach and provided comments at all the stages of the simulation and experimental works. Peter Sergeant and Essam Rashad revised the manuscript. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2. 3.

4.

5. 6. 7. 8. 9. 10.

11. 12. 13. 14. 15. 16. 17.

Mercorelli, P. A hysteresis hybrid extended kalman filter as an observer for sensorless valve control in camless internal combustion engines. IEEE Trans. Ind. Appl. 2012, 48, 1940–1949. [CrossRef] Mercorelli, P. A two-stage augmented extended kalman filter as an observer for sensorless valve control in camless internal combustion engines. IEEE Trans. Ind. Electron. 2012, 59, 4236–4247. [CrossRef] Bianchi, N.; Bolognani, S.; Bon, D.; Dai Pre, M.D. Rotor flux-barrier design for torque ripple reduction in synchronous reluctance and PM-assisted synchronous reluctance motors. IEEE Trans. Ind. Appl. 2009, 45, 921–928. [CrossRef] Nabil, M.; Allam, S.M.; Rashad, E.M. Modeling and design considerations of a photovoltaic energy source feeding a synchronous reluctance motor suitable for pumping systems. Ain Shams Eng. J. 2012, 3, 375–382. [CrossRef] Nabil, M.; Allam, S.M.; Rashad, E.M. Performance improvement of a photovoltaic pumping system using a synchronous reluctance motor. Electr. Power Compon. Syst. J. 2013, 41, 447–464. [CrossRef] Ibrahim, M.N.; Sergeant, P.; Rashad, E.M. Simple Design Approach for Low Torque Ripple and High Output Torque Synchronous Reluctance. Energies 2015, 9, 1–14. [CrossRef] Liu, C.T.; Chung, H.Y.; Lin, S.Y. On the electromagnetic steel selections and performance impact assessments of synchronous reluctance motors. IEEE Trans. Ind. Appl. 2017, 53, 2569–2577. [CrossRef] Ibrahim, M.N.; Sergeant, P.; Rashad, E.M. Synchronous reluctance motors performance based on different electrical steel grades. IEEE Trans. Magn. 2015, 51. [CrossRef] Morimoto, S.; Ooi, S.; Inoue, Y.; Sanada, M. Experimental evaluation of a rare-earth-free PMASynRM with ferrite magnets for automotive applications. IEEE Trans. Ind. Electron. 2014, 61, 5749–5756. [CrossRef] Barcaro, M.; Bianchi, N.; Magnussen, F. Permanent-magnet optimization in permanent-magnet-assisted synchronous reluctance motor for a wide constant-power speed range. IEEE Trans. Ind. Electron. 2012, 59, 2495–2502. [CrossRef] Spargo, C.M.; Mecrow, B.C.; Widmer, J.D.; Morton, C. Application of fractional-slot concentrated windings to synchronous reluctance motors. IEEE Trans. Ind. Appl. 2015, 51, 1446–1455. [CrossRef] Ibrahim, N.; Sergeant, P.; Rashad, E.M. Combined star-delta windings to improve synchronous reluctance motor performance. IEEE Trans. Energy Convers. 2016, 31, 1479–1487. [CrossRef] Ferrari, M.; Bianchi, N.; Doria, A.; Fornasiero, E. Design of synchronous reluctance motor for hybrid electric vehicles. IEEE Trans. Ind. Appl. 2015, 51, 3030–3040. [CrossRef] Reddy, P.B.; El-Refaie, A.M.; Galioto, S.; Alexander, J.P. Design of synchronous reluctance motor utilizing dual-phase material for traction applications. IEEE Trans. Ind. Appl. 2017, 53, 1948–1957. [CrossRef] Degano, M.; Carraro, E.; Bianchi, N. Selection criteria and robust optimization of a traction PM-assisted synchronous reluctance motor. IEEE Trans. Ind. Appl. 2015, 51, 4383–4391. [CrossRef] Bianchi, N.; Fornasiero, E.; Ferrari, M.; Castiello, M. Experimental comparison of PM-assisted synchronous reluctance motors. IEEE Trans. Ind. Appl. 2016, 52, 163–171. [CrossRef] Ibrahim, M.N.; Sergeant, P.; Rashad, E.M. Influence of rotor flux-barrier geometry on torque and torque ripple of permanent-magnet-assisted synchronous reluctance motors. In Proceedings of the 2016 XXII International Conference on Electrical Machines (ICEM), Lausanne, Switzerland, 4–7 September 2016; pp. 398–404.

Energies 2017, 10, 1500

18.

19. 20. 21. 22.

23.

18 of 18

Ibrahim, M.N.; Sergeant, P.; Rashad, E.M. Rotor design with and without permanent magnets and performance evaluation of synchronous reluctance motors. In Proceedings of the 2016 19th International Conference on Electrical Machines and Systems (ICEMS), Chiba, Japan, 13–16 November 2016; pp. 1–7. Chen, J.Y.; Chen, C.Z. Investigation of a new AC electrical machine winding. IEE Proc. Electr. Power Appl. 1998, 145, 125–132. [CrossRef] Lei, Y.; Zhao, Z.; Wang, S.; Dorrell, D.G.; Xu, W. Design and analysis of star–delta hybrid windings for high-voltage induction motors. IEEE Trans. Ind. Electronic. 2011, 58, 3758–3767. [CrossRef] Cistelecan, M.V.; Ferreira, F.J.T.E.; Popescu, M. Adjustable flux three-phase ac machines with combined multiple-step star–delta winding connections. IEEE Trans. Energy Convers. 2010, 25, 348–355. [CrossRef] Ibrahim, M.N.; Sergeant, P.; Rashad, E.M. Relevance of including saturation and position dependence in the inductances for accurate dynamic modelling and control of SynRMs. IEEE Trans. Ind. Appl. 2017, 53, 151–160. [CrossRef] De Almeida, T.; Ferreira, F.J.T.E.; Duarte, A.Q. Technical and economical considerations on super high efficiency three-phase motors. IEEE. Trans. Ind. Appl. 2014, 50, 1274–1285. [CrossRef] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).