A Novel Switched Reluctance Motor With C-Core Stators

Download as pdf or txt
Download as pdf or txt
You are on page 1of 8
 
IEEE TRANSACTIONS ON MAGNETICS, VOL. 41, NO. 12, DECEMBER 2005 4413
A Novel Switched Reluctance MotorWith C-Core Stators
Shang-Hsun Mao and Mi-Ching Tsai
Department of Mechanical Engineering, National Cheng Kung University, Tainan 701, Taiwan, R.O.C.
This paper presents a novel switched reluctance motor (SRM) design in which the stator is simply formed from C-cores. Unlike con-ventional SRMs, the windings of the new motor can be individually wound into the stator cores without complex winding equipment.Because of the inherent axial fielddistribution, this typeof SRM requires a three-dimensional (3-D) finite-element analysis (FEA) modelfordetailedfluxanalysis.Thispaperproposesanapproximatedtwo-dimensionalFEAmodeltospeedupcomputationaltime.Inaddition,since the proper current that ensures operation in the saturated region (to maximize torque and efficiency) is often hard to determinesystemically, the paper proposes a simple method to determine the optimum operating current. Finally, the paper compares some char-acteristics of a traditional SRM with those of the proposed SRM. The comparison shows that the proposed SRM performs well in termsof torque and efficiency, and provides a higher degree of flexibility in winding design.
 Index Terms—
Finite-element analysis, optimum operating current, switched reluctance motor.
I. I
NTRODUCTION
S
WITCHED reluctance motors (SRMs) have been widelyused in many industrial applications such as aerospace, au-tomotive and domestic appliance manufacturing [1]. The majoradvantages of SRMsare hightorque output,widerangeof oper-atingspeed, simplestructureandfault tolerance.Severalanalyt-ical methods for SRM design have been presented [2], [3]. Be- cause of its inherent nonlinear magnetic characteristics and thedoubly salient pole structure, the finite element analysis (FEA)approach is often adopted for obtaining its accurate magneticcharacteristics [4], [5]. TwoofthemajordesignobjectiveswhendesigningSRMsarehigh efficiency and low cost. Fig. 1 shows the basic structureof a conventional SRM. While the geometry and windings arenot as complex as, say, an induction motor, automated windingof this machine can still be a complex process. Also the per-formance can be limited by the area of the slots and the tight-ness of winding (i.e., the slot-fill and end-winding overhang).For example, if the winding is not tight enough, the end-turnsof winding will cause large copper loss and then the motorefficiency will be reduced. There will also be additional end-winding resistance and leakage inductance which will increasethedemandonthesupply.ThispaperpresentsanovelSRMwitha C-core type of stator for axial winding which is based on [6].Such a structure creates a large slot space so that the windingprocess becomes more straightforward. Also, since the C-corestators are independent of each other, they can be manufacturedandwoundindividuallyandefficientlybyanautomatedprocess.Consequently the cost of winding will be reduced significantly.However, the flux distribution of this new type of SRM is inher-ently three-dimensional (3-D) and hence requires a 3-D finiteelement model for analysis. It is well known that 3-D analysisis very time consuming in simulation so in this paper, a simpli-fiedtwo-dimensional(2-D)analysisoftheoriginal3-Dmodelis
Digital Object Identifier 10.1109/TMAG.2005.858372
proposed in order to speed up the simulation process. The An-soft/Maxwell finite element analysis package [7] is employedfortherequiredsimulationandalsotocalculatethefluxlinkage,inductance and torque.In general, the output torque of an SRM is related to thechange of co-energy so it should be designed to operate intothe saturated region for greater output torque. However, theproblem is how to determine the optimum operating current. Inthis paper, a method is developed to determine the operatingcurrent and turns-per-phase, where the flux linkage is modeledwith a single curve in the unaligned position and two curvesin the aligned position respectively [8]. By calculating themaximum increment of co-energy, the optimum excitationcurrent that produces the maximum increment of torque can bedetermined.This paper is organized as follows. Section II presents the ge-ometry design based on the “feasible” triangle criteria [9]. Asimplified 2-D model is proposed in Section III. Determinationof number of the coil turns and the rated current is given in Sec-tion IV. The characteristics comparison with a traditional SRMis shown in Section V. Section VI concludes the paper.II. M
OTOR
 D
ESCRIPTION AND
 G
EOMETRY
 D
ESIGN
 A. Motor Description
The structure of traditional SRM has a radial winding andradial air gap. Let the electric loading be defined as [10](1)where is the number of coil-turns per phase, is the phasecurrent, is the number of phases conducting simultaneously,and is the bore diameter. It can be seen that the product of and isaconstantforagivenelectricloadingandborediameter.However, the winding space is restrained by slot space, numberof turns, area of conductor cross section, insulation thicknessand the maximum manufacturing slot-fill. Obviously, it is dif-ficult to determine the number of turns and phase current in aconfined slot and the design flexibility will be limited by the
0018-9464/$20.00 © 2005 IEEE
 
4414 IEEE TRANSACTIONS ON MAGNETICS, VOL. 41, NO. 12, DECEMBER 2005
Fig. 1. Conventional four-phase 8/6 SRM.
slot space. Note that the motor performance can be restricted bythe slot space, especially in the SRM.Consider the novel SRM shown in Fig. 2(a), where thestator is composed of eight independent C-cores which arearranged toroidally along the shaft with the windings locatedin the middle of the C-cores. Such an SRM has the followingpotential advantages.1) Compared with the conventional SRM, the slot space of the novel SRM is relatively large so that there is more de-sign
 fl
exibility for the number of turns and phase magne-tomotive force (MMF).2) The stator winding can be wound more closely and uni-formly so that the end turn of the winding is shortened.This helps to reduce the copper waste and copper loss.Also, the end-winding
 fl
ux can be reduced.3) Therotor inertiaof theproposed SRM canbe made small,which is suitable for high-speed applications.4) Such a structure can easily be augmented by cascadingseveral modules on one shaft for high power operation asshown in Fig. 2(b), which illustrates two modules.5) Thestatorsareindependent,andcanbewoundindividuallyandautomatically.FromthesideviewasshowninFig.3(a),the material in the dashed blocks can be saved for theproposed SRM, and from the top view as shown inFig. 3(b), there is no back iron (the dashed block) in theproposedSRMcomparedtothetraditionalone.Therefore,material waste and manufacturing cost can be reducedsigni
cantly.
 B. Geometry Design
The
uxpathandthegeometryparametersforthenewdesignofSRMareillustratedinFig.4,wherethetwomajorparameterswhich greatly in
uence the inductance curves are the rotor polearc, , and the stator pole arc, . The design of the pole arcsshould satisfy the basic requirements for self-starting, highertorque output, and lower torque ripple.
Fig. 2. (a) Structure of the proposed SRM. (b) Two modular expansion of theSRM.
To achieve the self-starting requirement, must be greaterthan the stroke , which is de
ned by(2)i.e.,(3)where is the number of phases and is the number of rotorpoles. For example, the stroke of a four-phase 8/6 SRM is 15 .Tokeepthealignedtounalignedinductanceratioashighaspos-sible(tomaximizethecurrent/ 
uxlinkageloopforhightorque),the stator pole must be smaller than the interpolar arc of therotor, i.e.,(4)Furthermore, to reduce torque ripple, the positive slope of theinductancecurvemustoverlapthenegativeslopeofthepreviousphase. We thus require that the rotor pole arc should be slightlygreater than the stator pole arc so that(5)The above constraints are summarized and illustrated inFig. 5. The possible combinations of the rotor pole arc andstator pole arc are located in the
 “
feasible
 triangle.The basic dimensions are: a rotor pole arc of 24 , a statorpole arc of 18 , and an air-gap length of 0.5 mm. For ease of manufacture, the radius of the shaft is chosen to be 4 mm,the outer radius of the rotor mm, and the length othe stator pole mm. To ensure the
 fl
ux that can passthroughfromthestatortotherotor,boththewidthofstatoryokeandthedepthofrotorpole aresettobeequaltothelengthof the stator pole. Now considering the slot spacing, the lengthof stator is then chosen to be 70 mm. To sum up the abovedesign, the geometric parameters, which are summarized as inTable I, appear to be reasonable for the constraints previouslydescribed.III. 2-D M
ODIFIED
 M
ODEL
With the given geometric parameters above, a 3-D FEAmodel for the prototype SRM was developed as shown in
Scribd
You're Reading a Preview

Upload your documents to download.

Or

Become a Scribd member for full access. Your first 30 days are free.

 
MAO AND TSAI: A NOVEL SWITCHED RELUCTANCE MOTOR WITH C-CORE STATORS 4415
Fig. 3. (a) Side view of the proposed SRM. (b) Top view.Fig. 4. Flux path and the geometry parameters of the proposed SRM.
Fig. 6. The
 fl
ux distributions at the aligned and unaligned po-sitions are shown in Fig. 7, where only one phase is illustratedand switched on. Obviously, the axial
 fl
ux in the stator and theradial
 
ux in the rotor appear in the 3-D
 fl
ux distribution andthere is a long axial
 fl
ux path in the stator. This implies a 3-Dmodel is required for magnetic
 fi
eld simulation in order to takeinto account all of the 3-D effects. However, it is well knownthat the use of 3-D FEA often has the following disadvantages[11].1) The problem de
nitions such as model drawing, ma-terial assignment, and boundary condition can be verycomplicated.2) The solution may not always converge because the meshof the air gap is hard to re
ne. The mesh of the air gapused for an SRM simulation is an important factor andcan seriously affect the accuracy of the entire solution.3) A 3-D FEA model that uses a large number of meshes toobtain an accurate solution requires a very long computa-tion time
both in setting up the model and running theanalysis.To improve the simulation ef 
ciency, this paper presents asimpli
ed 2-D FEA model as shown in Fig. 8 to achieve therequired simulation; where two stators are modeled in the sameplane. The machine is rearranged as illustrated. The aligned and
Scribd
You're Reading a Preview

Upload your documents to download.

Or

Become a Scribd member for full access. Your first 30 days are free.

 
4416 IEEE TRANSACTIONS ON MAGNETICS, VOL. 41, NO. 12, DECEMBER 2005
Fig. 5.
 
Feasible
 triangle for four-phase 8/6 SRM.TABLE IG
EOMETRY
 P
ARAMETERS OF THE
 P
ROPOSED
 SRMFig. 6. 3-D FEA model.
unaligned positions can be expressed graphically as shown inFig. 8(a) and (b). It should be noticed that the width of the statoryoke in Fig. 8 is equal to the width of the stator pole in Fig. 4,makingthecross-sectionalareaofthestatorcore-back constant,since the stack length of the 2-D model is equal to the length of the stator pole in Fig. 4. The advantages of the proposed 2-DFEA model are as follows.1) It is simple to build using a 2-D FEA model, includingmodel geometry, material assignment, and boundaryconditions.2) The mesh of the air gap is easy to re
ne for a more accu-rate solution. Note that most of magnetic energy is storedin the air gap, so that the mesh of the air gap plays a sig-ni
cant role in the simulation.3) Thesimpli
ed2-DFEAmodelresultsinagreatreductionof simulation time.4) Most characteristics can be determined from the 2-Dmodel such as
 
ux linkage curves, inductance curves,and static torque.In this paper, a 3-D model simulation is
 fi
nally employed toverifythe accuracyof the2-D simpli
ed model. The inductancecurves obtained from the 2-D and 3-D simulations with respecttotherotor position are shownin Fig.9whiletheexcitationcur-rent was 500 A. As can be seen, the inductance curve resultsfrom the 2-D simulation shows good correlation with resultsfrom the 3-D simulation. It should be noted that the 3-D sim-ulation takes several days to calculate for the inductance curve.However, the inductance curve can be obtained in few hoursusing the proposed simpli
ed 2-D inductance simulation. Thisvalidates the method although it should be borne in mind thatthe stator-yoke leakage differences should be carefully consid-ered so that the 2-D and 3-D values approximately match.The
 fl
ux linkage curves of the aligned position and unalignedposition with respect to the current are shown in Fig. 10 andare calculated using the 2-D FEA model. This determines theoperating current and turns-per-phase as described below.IV. O
PERATING
 C
URRENT AND
 N
UMBER OF
 T
URNS
P
ER
 P
HASE
 D
ETERMINATION
Let the co-energy be de
ned by(6)The average torque of the SRM can be computed as [12](7)where is the variation of co-energy over one stroke. Thisshows that the larger change of co-energy, the higher increasein average torque is. Nevertheless, the enhancement of outputtorque is not uniform, especially in the saturated region. Theo-retically, the maximum increment of co-energy and torque willoccur at the same current. Therefore by
 fi
nding out the currentwhich produces a maximum increment of co-energy the max-imum increment of torque is also obtained since it will be pro-duced at the same current. The maximum increment of co-en-ergyisthenanindexfordeterminingthecurrentandthenumberof turns-per-phase. In order to obtain the variation of co-energy,the modeling of the
 fl
ux linkage curve is necessary.For the aligned position, both linear and nonlinear curves areused to approximate the
 fl
ux linkage curve as shown in Fig. 11where the linear curve 1 is modeled as a straight line and the
Scribd
You're Reading a Preview

Upload your documents to download.

Or

Become a Scribd member for full access. Your first 30 days are free.

Reward Your Curiosity

Everything you want to read.
Anytime. Anywhere. Any device.
No Commitment. Cancel anytime.
576648e32a3d8b82ca71961b7a986505