Winding Arrangement of a New Type Hollow Rotor BLDC Motor

Received Mar 15, 2018 Revised Mar 22, 2018 Accepted Jul 1, 2018 This paper discusses about winding arrangement of fractional slot of a new type hollow rotor Brushless Direct Current (BLDC) motor. Hollow rotor has higher performance compared to other BLDC motor because it minimizes the unused flux below permanent magnet and maximize torque produce by the motor. It’s also known that 8 pole motor is favorite used in industrial because it has an optimum space of permanent magnet for a smaller motor size. The number of pole will affect the maximum speed of the rotor. Thus, the objective of this research is to investigate the best winding arrangement for 8 pole of hollow rotor that could produce the highest electromagnetic performance. At starts, four combinations of slot number and coil sizes had been selected. Structural comparison in term of coil vector and winding arrangement is studied. Finite Element Method (FEM) had been used to simulate the parameters such as backemf and torque waveforms. It was convinced that 9 slot 8 pole with 0.6 size of coil produces the best performance. The confirmed model had been fabricated and measured. Both results from FEM and measurement are compared in term of backemf and torque where percentage differences are 7.4 % and 8 %, respectively. As conclusion, this research shows the fundamental of winding arrangement of fractional slot of motor especially 8 pole motor. Keyword:


INTRODUCTION
The Brushlesss Direct Current (BLDC) motor is widely used in applications including appliances, automotive, aerospace, consumer, medical, automated industrial equipment and instrumentation. The BLDC motor is electrically commutated by power switches instead of brushes. Compared with a brushed DC motor or an induction motor, the BLDC motor has many advantages such as higher efficiency and reliability, lower acoustic noise, smaller and lighter, greater dynamic response, better speed versus torque characteristics, higher speed range and longer life. For instance, they own the highest torque density and efficiency among all types of motors. Moreover, thanks to their flux weakening capability, interior BLDC allow wide speed operation with constant power making them suitable for traction and transportation systems [1]. It is very important to pay high attention on combination of pole and slot numbers. Taeyong Yoon studied about magnetically induced vibration in a permanent-magnet BLDC with symmetric pole-slot configuration [2]. The author reports a numerical and experimental study of magnetically induced vibration associated with rotor and stator eccentricity and imperfect magnetization for 8-pole 6-slot symmetric BLDC motors. The results show that there is an optimal slot angle for minimum cogging torque, but this slot angle is not optimal for reducing magnetic forces. Many literatures has presented improvement methods for the reduction of

BASIC STRUCTURE
The basic structure of conventional BLDC motor is shown in Figure 1 (a). Rotor is covered by silicon steel material, it will influence flux of permanent magnet to pass through the area below the ferromagnetic material. As a result, there is more leakage flux circling at the end of the permanent magnet as shown in Figure 1 (c). To reduce leakage at the end of permanent magnet, area below rotor is changed to hollow rotor as shown in Figure 1 (b). The arrangement of magnet in a hollow rotor is different with other BLDC motor where magnet is arranged to maximally reduce leakage flux. Hollow which consists of an air, has higher reluctance, thus channeling all flux to the nearest ferromagnetic material. In a hollow rotor spoke type BLDC motor, all flux will go around stator where it will maximize the usage of flux in the motor. Figure  1 (d) shows generated flux for hollow-rotor BLDC motor. Hollow which consists of an air, has higher reluctance, thus channeling all flux to the nearest ferromagnetic material. In a hollow rotor spoke type BLDC motor, all flux will go around stator where it will maximize the usage of flux in the motor. Table 1 shows basic structure configuration for all combination and magnetic flux. Slot and pole parameter for each design is tabulated in Table 2. Each structure has the same value of rotor inner diameter, r1, rotor outer diameter, r2, stator inner diameter, r3, stator outer diameter, r4, stator tooth height, s1, stator width height, s2, air gap, ag, and permanent magnet volume which are 20 mm, 24.8 mm, 47 mm, 50 mm, 3 mm, 1.5 mm, 0.2, and 320 mm2, respectively. Pole for each design is 8. The different between each design is only slot number. Varied parameter for each design is the number of turn and current where the number of turn is varied from 18 to 50 while current is varied from 2 A to 10 A. Flux lines for all design is tabulated from 0 T to 2 T which is maximum value. All the flux for all design is recorded during backemf

. Pitch factor and distribution factor in motor windings
For fractional slot and pole number motors, the slot pitch, sp and pole pitch, pp are generally given by equation (1) and equation (2) where n is odd harmonics orders due to symmetrical permanent magnet distribution in the airgap region. Referring to Figure 2 (a), taking 6 slot 8 pole as an example, the coil is short chorded by a pitch angle αp as in equation (3) where Ns is number of slot and P is pole of motor [25]. The pitch factor and distribution factor can be calculated using Z. Q. Zhu, M. S. Ahmad and I. Dahaman method [26]- [29] as shown below.
(1) Both sides of the coil will have an induced backemf, but because their angle is not zero, the backemf E1 and E2 are added vectorially to yield the coil backemf Er, whose magnitude is influenced by the pitch factor, defined as the ratio of the resultant vector, Er to the algebraic sum of the individual vectors E1 and E2 as shown in Figure 2 (b). Therefore, the pitch factor, Kpn for 6 slot 8 pole is given by equation (4) For 6 slot 8 pole motor, each phase winding has 2 coil connected in series. Individually, each coil produces backemf vector as shown in Figure 3 (a). The combination of both coils to form phase winding which leads to highest distribution factor, kdn will be chosen. Figure 3 (b) and Figure 3 (c) shows selection of coil per phase and derivation of distribution factor, respectively. The equation can be described as shown in equation (5) to (13).
Resultant emf vector per phase: The distribution factor for 6 slot 8 pole, kdn can be shown as: Hence, the winding factor Kdpn for 6 slot 8 pole is the product of the pitch factor and distribution factor such that: Similar approach and method can be used for 9 slot 8 pole and 12 slot 8 pole motor. For 9 slot 8 pole, each winding consists of three coils connected in series. Figure 4 (a) shows how individual coils are selected to form the phase winding while Figure 4 (b) shows selection of coil for each phase. Coil 3, 8 and 4 are selected to form phase A windings. Polarities for the connection of coil 3 and 4, however have to be reversed in order to generate maximum backemf. Figure 4 (c) shows derivation of distribution factors. The resultant emf vector per phase is given by equation (14): Derivation of distribution factors for 9 slot 8 pole: The derivation factor can be derived as equation (15) and (17): Hence the winding factor for 9 slot 8 pole for the motor is given by equation (18) to (20): For 12 slot 8 pole, each phase winding has 4 coils connected in series. Individually, each coil produces backemf vector as shown in Figure 5 (a). The combination of both coils to form phase winding which leads to highest distribution factor, kdn will be chosen. Figure 5 (b) and Figure 5 (c) shows selection of coil per phase and derivation of distribution factor, respectively. The derivation factor for 12 slot 8 pole can be derived as shown in equation (21) and (23).   Table 3 shows coil vector and winding arrangement for each phase which is separated between dotted line for all hollow rotor slot and combination. Each arrangement is design for three phase configuration. For 6 slot 8 pole each phase has two winding.

PARAMETER FOR ANALYSIS
This research is aimed to study the best possible combination of slot number for 8 pole hollow rotor. The general methodology is shown in Figure 6 (a). Firstly, four combination of slot and pole is selected which are 6 slot 8 pole, 9 slot 8 pole, 12 slot 8 pole and 15 slot 8 pole. All the parameter for design in term of rotor inner diameter, r1, rotor outer diameter, r2, stator inner diameter, r3, stator outer diameter, r4, stator tooth height, s1 stator width height, s2, air gap, ag, and permanent magnet volume is fixed. Only winding arrangement of the motor is different. The design carried out by using computer aided software (CAD) software. Coil vector and winding arrangement of each combination is properly designed. Analysis parameter for all design in term of backemf, stator flux density, rotor flux density, static torque and transient torque is simulated by using FEM analysis. During backemf analysis, diameter of coil is varied from 0.4 mm to 1.0 mm. One design from all four combination will be choosen and fabricated based on the best motor performance. The Finite Element Method (FEM) starts with the developing of the structure modelling of the electric machine as shown in Figure 6 (b). For structure modelling, user is required to sketch the geometry in (CAD) and imports the geometry files in the FEM software. Meshing calculation for the model will be carried out before element calculation took place. It requires time to develop the model depends on the model structure. When mesh development is completed, system will start element calculation. At this point, all the calculation for FEM will be conducted. After that, declaration of material and configuration will be carried out. During this stage, setup for material and other configuration such as simulation time, air gap setting, motion setup and boundaries will be selected. Magnetic analysis will be simulated and result for each analysis will be recorded. There are some parameter which will change during FEM calculation. Figure 7 shows the simulation result using FEM analysis for backemf. Maximum speed for the motor is fixed for 1000 rpm. For all design, maximum backemf occur when coil size of 0.4 mm is used while minimum backemf occur when coil size of 1.0 mm is used. This is because, smaller diameter coil size will have higher number of turn. Figure 8 shows maximum torque analysis by using FEM. All torque result is captured at speed 1000 rpm and current range from 2 A to 10 A. Figure 8 Figure 9 shows flux density for stator and rotor. At this point, current is applied from range 2 A to 10 A. The flux produce include magnetic flux from permanent magnet and current. The flux density is amount of flux through an area as shown in equation (17) where B is flux density, ϕ is flux and A is area of passing flux. Figure 9 is result of stator flux density for all slot and pole combination for coil diameter range from 0.4 mm to 1.0 mm. Stator flux density and rotor flux density is defined at maximum speed 1000 rpm. Figure  9 Figure 9 (b) shows rotor flux density for coil diameter size of 0.4 mm. Maximum flux density is 1.46 T for 12 slot 8 pole arrangement while minimum flux density for rotor is 0.75 T for 6 slot 8 pole arrangement. The 12 slot 8 pole and 9 slot 8 pole arrangement has the most nearest value of flux density which the percentage different is between 33.3 % to 11.23 %. Figure 9 (c) shows stator flux density for coil diameter 0.6 mm. The maximum stator flux density appears for 12 slot 8 pole design which is 1.28 T while minimum flux density is 1.19 T for 9 slot 8 pole arrangement. For this coil size, 9 slot 8 pole has the lowest stator flux density when the current is change from 2 A to 10 A where the maximum flux density at this point is 1.62 T. Figure 9 (d) shows result of rotor flux density for all slot and pole combination. Result for rotor flux density is investigated for current 2 A to 10 A at coil size 0.6 mm. Maximum rotor flux density occur at current 10 A for design 15 slot 8 pole which is 1.15 T while minimum rotor flux density is 0.58 T at current 6 A for 6 slot 8 pole design. Rotor flux density increase linearly with the increase of current for all design. The highest rotor flux density appear for 12 slot 8 pole while lowest rotor flux density occur for 6 slot 8 pole design. Figure 9 (e) shows stator flux density for coil size of 0.8 mm where maximum flux density is 1.71 T for 15 slot 8 pole and 12 slot 8 pole arrangement. Figure 9 (f) shows rotor flux density result for coil diameter size of 0.8 mm. The maximum flux density is 1.08 T for 15 slot 8 pole and 12 slot 8 pole arrangement. The minimum rotor flux density is 0.46 T for 9 slot 8 pole arrangement. Figure 9 (g) shows stator flux density result for coil size of 1.0 mm. The maximum stator flux density is 2.07 T for 15 slot 8 pole combination while minimum stator flux density is 0.95 T for 9 slot 8 pole arrangement. It can be seen from the graph stator flux density for 6 slot 8 pole and 12 slot 8 pole arrangement has only small different which is less than 5 %. Figure 9 (h) shows the result for rotor flux density when 1.0 mm coil size is used. The maximum flux density is 1.35 T for 15 slot 8 pole design while minimum flux density is 0.35 for 9 slot 8 pole arrangement. The minimum stator flux density is 0.95 T for 9 slot 8 pole arrangement. From result 9 (a), 9 (c), 9 (e) and 9 (g), it can be seen that 9 slot 8 pole arrangement has the lowest stator flux density which makes it on the top reasonable design among all other combination. When it is compared to rotor flux density, 9 slot 8 pole design has the range of flux density below 1.5 T. From all the result from Figure (8) and Figure (9), it is concluded that coil size of 0.6 mm is the best coil size used for motor prototype. This is because, when bigger coil size is used, lower torque is obtained while when smaller coil size is used, higher flux density will be produced. This will lowered the motor performance where higher flux density will bring motor performance to be saturated. The complete prototype is shown in Figure 10 (d). Casing length, cl, casing width, cw, and casing height, ch is set to 50 mm, 50 mm and 10 mm, respectively. Stator and rotor is made from laminated steel. The fabricated hollow rotor is 9 slot 8 poles with the winding of coil diameter size 0.6 mm. The heat produce by coil during motor operation will be dissipated to surrounding air. There are two casing used which will cover front and back of the motor. To fit both casing with the stator, four set of screw will be used at each end of the casing. Material used for silicon steel is standard silicon J1: 50800, permanent magnet is Neodynium Nd 42 and copper for coil. Table 5 shows hollow rotor specification for the fabricated motor. Casing length, cl, casing width, cw and casing height, ch is 50 mm, 50 mm and 10 mm, respectively. Number of slot and pole for the for the fabricated motor is 9 slots 8 pole. The permanent magnet volume is 320 mm. Maximum current for hollow rotor operation is 6 A. Number of turn set for each winding is 30. Air gap, ag is 0.5 mm.   Figure 11 (a) shows diagram for torque and speed measurement setup. Speed sensor will measure the speed in rpm which will be shown by digital speed indicator. Torque sensor is connected to dynamic strain amplifier which allows the data to be presented in oscilloscope. For this experiment, a BLDC driver is used to operate hollow rotor BLDC motor. 15 V is being supply to the BLDC driver to trigging the switching device (MOSFET). 5 V is set for supplying hall effect sensor, meanwhile the motor supply is set for 12 V initially. Then, for next experiment the motor supply is varied to 24 V and 48 V. Three phase wye connection is set to BLDC driver. Powder brake is connected to torque sensor shaft. Other power supply is used for powder brake with 0.5 V, 1.0 V, 1.5 V and 2.0 V to provide different value of braking torque. Figure 11 Figure 12 shows verification of hollow rotor by using FEM simulation and measurement. The current is fixed 6 A. Figure 12 (a) shows hollow rotor back emf characteristic. Maximum back emf using FEM is 5.4 V while measurement result is 5 V. Percentage different is 7.4 % between simulation and measurement. For the measurement of back emf, a constant speed of 1000 rpm is given to the motor for the recording purpose. Only phase A is compared for FEM simulation and measurement because all the three phases have the same pattern but with different phases. Figure 12 (b) shows maximum backemf for measurement and FEM. Maximum speed for backemf is 1000 rpm. Minimum backemf at 200 rpm during FEM and measurement is 1.14 V and 1.026 V, respectively. Meanwhile maximum backemf is 5.54 V and 4.986 V for measuremet and FEM at 1000 rpm. Backemf increase linearly with speed. Percentage different between FEM and measuremet is 10 %. Figure 12 (c) shows the comparison result for static torque characteristic. The maximum value of torque during measurement is 0.31 Nm while the maximum value of torque using FEM simulation is 0.39 Nm. The percentage error between measurement and simulation is about 8 %. During measurement of static torque, motor will be in static position where motor will be rotating without the use of any driver. Static torque is compared with simulation when pure DC current source is given in phase winding. Dynamic torque is shown in Figure 12 (d). Maximum dynamic torque from measurement is 0.31 Nm at current 5.4 A, while from FEM is 0.34 Nm. The average value of dynamic torque from FEM compared to measurement has percentage of difference is 4 %. As a conclusion, all result has good agreement between simulation and measurement. Torque ripple measures as the difference in maximum torque and minimum torque over average torque of one revolution as shown in equation (18).

VERIFICATION RESULT
[%] (18) Whereas, Tmax is the maximum torque, Tmin is the minimum torque while Tave is the average torque. Basically, torque ripple is generated by the pulsation of the cogging torque, the variation of induction with the rotor position, and the interaction between the magnet-motive force of stator and the magnetic force of the permanent magnet. The result for torque ripple is tabulated as in Table 6.

CONCLUSION
In this paper, winding arrangement of new type hollow rotor had been discussed. Hollow rotor has higher performance compared to other BLDC motor design because it minimize the unused flux below permanent magnet and maximize torque produce by the motor. Four winding arrangement is selected and design process for hollow rotor is carried out by using desired parameter. Structural comparison in term of coil vector and winding arrangement is analyzed by using FEM. Four sizes of coil diameter had been selected in order to select the best coil size. 9 slot 8 pole design with 0.6 mm size of coil diameter is selected for fabrication. Measurement and FEM result is compared in term of backemf and torque where it has percentage differences of 7.4% and 8%, respectively.