Design and Performance Analysis of an Axial-Flux Disk-Type Switched Reluctance Motor for Hybrid Scooters

882 Design and Performance Analysis of an Axial-Flux Disk-Type Switched Reluctance Motor for Hybrid Scooters Jian-Feng TSAI and Yon-Ping CHEN Switched reluctance motors (SRM) possess several superiorities such as simplicity in structure, ruggedness, and good reliability, which make it viable for the transportation applications. In this paper, an axial-flux disk-type switched reluctance motor (DSRM) for hybrid scooters is proposed. The DSRM can be used to drive the scooter and as an engine starter simultaneously. Moreover, the torque density of axial flux motors is more superior to that of radial flux motors in this application. Instead of utilizing FEM, the permeance model is employed to model the DSRM. With the analytical expressions, the output torque and current can be obtained dynamically and the geometries of the DSRM can be further optimized. Finally, the hardware-in-the-loop (HIL) simulation environment is constructed to perform full system simulation and the DSRM is shown to be capable of being an auxiliary power source for hybrid scooters. Key Words: Switched Reluctance Motor (SRM), Axial-Flux Motor, Hardware-in-the-Loop (HIL), Hybrid Scooter 1. Introduction Due to the high emissions and low fuel economy of gasoline scooters in urban areas, a variety of electric scooters have been proposed as the substitutes in many literatures. However, electric scooters have to encounter the problem of low acceleration and short cruising time. In order to overcome the above drawbacks, one alternating choice is the hybrid scooter, which consists of a gasoline engine and an additional electric motor. There are many different kinds of electric motors used for transportation applications, such as permanent magnet (PM) motors, induction motors, switched reluctance motors (SRMs), etc. Compared with PM motors and induction motors, the SRMs possess several outstanding characteristics such as simplicity in mechanical structure, ruggedness, good reliability and lower hysteresis loss (1). With these advantages, the SRMs gain more attention recently and have been treated as a better choice for the transportation applications (2) (4). Generally, from the difference in flux distribution, the electrical machines can be classified into radial-flux machine (RFM) and axial-flux machine (AFM). The RFM Received 6th May, 2005 (No. 05-5039) Department of Electrical and Control Engineering, National Chiao Tung University, 1001 Ta Hsueh Road, Hsinchu 300, R.O.C. E-mail: ypchen@cc.nctu.edu.tw Series C, Vol. 49, No. 3, 2006 has been popularly used in many industrial applications and also in transportation. However, with essential distinction in flux distribution, the AFM will be a better choice than RFM for some applications. Huang et al. used the sizing equations to compare the power density of the RFM and AFM (5). Cavagnino et al. used simple thermal consideration to compare RFM and AFM with permanent magnet and concluded that if the ratio between the motor overall axial length and the motor external diameter is short, AFM will possess higher torque density than RFM (6). In this paper, an axial-flux disk-type switched reluctance motor (DSRM) is designed for modifying the gasoline scooter into a hybrid scooter. Due to the space limitation in the existing gasoline scooters, DSRM will provide higher torque density and can simplify the mechanism of transmission. In the following sections, the configurations of hybrid scooters will be discussed firstly. Then, the design and derivation of the analytical model for DSRM will be processed, and the characteristic profiles can be obtained afterward. Finally, full system simulation will be performed in the HIL simulation environment to validate the output performance of the DSRM. 2. Configuration of Hybrid Scooters Recently, electric scooters have been considered as a suitable substitutive for traditional gasoline scooters, JSME International Journal

883 Table 1 Comparison of the electric and hybrid scooters which produce tremendous harmful emissions. However, the use of electric scooters has to face two critical drawbacks, low acceleration and short cruising time. In order to reduce the emissions and keep the merits of the gasoline engine, investigators have paid their attentions to the design of hybrid scooters to meet the requirements, such as the 50 cc hybrid scooter prototyped by HONDA Corp. in August 2004. Several comparisons for the electric and hybrid scooters are listed in Table 1. A hybrid powertrain consists of two or more power sources to generate propulsion power. Generally, there are two types of powertrain configuration in hybrid scooters, serial and parallel. The main difference between them is how the propulsion power from gasoline engine and electric motor are distributed. The gasoline engine in serial hybrid scooters is used purely to drive the generator and then generate electricity, which is distributed through the power bus, partially for drive the motor and partially stored in the battery pack. Since the gasoline engine is not connected to the drive-wheel directly, it could always operate at highest efficiency. However, the peak power required for accelerating and cruising at high speed is limited by the capability of electric motor. As for parallel hybrid scooters, both the gasoline engine and the electric motor are coupled together to drive the driving-wheel. Because the power from the gasoline engine and electric motor is sent to the driving-wheel simultaneously, higher peak power will be obtained comparing to the serial configuration. However, an additional torque coupler is needed in this parallel configuration, such as the planetary gear set. Thus, this paper focuses on the designing of a parallel hybrid scooter with the DSRM, which has higher peak power and more simple mechanism. 3. Design of the DSRM The configuration of the parallel hybrid scooter with the DSRM is shown in Fig. 1. The DSRM is integrated into the gasoline scooter in the limited space between the engine and the continuous variable transmission (CVT) to simplify the torque coupler. 3. 1 The mechanism of DSRM As shown in Fig. 1, the symbol (1) denotes the shaft that connects the rotor of the DSRM with CVT and the symbol (2) denotes the plate that connects with the shaft Fig. 1 Fig. 2 The powertrain configuration of the parallel hybrid scooter with DSRM Structure and geometries of the 6/4DSRM of the engine. Once the clutch connects the rotors with (2), the shaft of the DSRM and the engine will be locked together. At low speed, only the DSRM drives the scooter, since the engine s efficiency is very low but with lots of emission. When the speed increasing, the clutch connects the DSRM with the gasoline engine and then the DSRM acts as an engine starter to activate the engine. Thereafter, the scooter may be driven by the engine only or simultaneously by the DSRM which assists the engine during the heavy load or high power demanded. Besides, the DSRM can be further used as a braking generator during decelerating (3). With this configuration and the DSRM, one can reform the existing gasoline scooter into a parallel hybrid scooter with minor modifications. 3. 2 The structure and geometry of DSRM The structure of the 6/4 DSRM (six stator poles and four rotor poles) is depicted in Fig. 2. There are three stacks in the 6/4 DSRM and each of them consists of rotor/stator poles and carriers. Unlike the PM motor, the rotor contains only iron laminations without any permanent magnet. Thus, the structure of the DSRM is very rugged JSME International Journal Series C, Vol. 49, No. 3, 2006

884 Table 2 Dimensions of the DSRM Table 3 Electrical parameters and inductance of the DSRM which makes it an appropriate choice for the transportation applications. The stator poles are concentrated-winding and could be energized without polarity. The drivers of each winding are all independent such that the DSRM is quite reliable. Moreover, the core losses can be reduced since the flux-paths are shorter than the conventional design. Detailed dimensions of the DSRM are shown in Fig. 2 and the named dimensions are listed in Table 2. 4. Analytical Model and Operation of the DSRM 4. 1 Analytical inductance model of the DSRM Due to the high nonlinearity in the DSRM, it is hard to model the inductance profile analytically with respect to the rotor position. To deal with the problem, Fahimi et al. used Fourier series to relate the self-inductance with the rotor position since the inductance is a periodic function of the rotor position (8). Although only first three coefficients of the Fourier series are considered, Fahimi et al. obtained quite accurate results close to the FEM analysis, which can be approximately expressed as L(θ) L 0 + L 1 cos(n r θ)+ L 2 cos(2n r θ) (1) where L 0, L 1, L 2 are the coefficients and θ is the angle of rotor. Then, the inductance profile of stator poles can be formulated as: L a (θ) = L a (θ) L 0 + L 1 cos(n r θ)+ L 2 cos(2n r θ) (2) L b (θ) = L b (θ) L 0 + L 1 cos(n r θ π/3)+ L 2 cos(2n r θ π/3) (3) L c (θ) = L c (θ) L 0 + L 1 cos(n r θ 2π/3)+ L 2 cos(2n r θ 2π/3) (4) where the lowercase denotes the six stator poles, phase-a, phase-a, phase-b, phase-b, phase-c and phase-c. Series C, Vol. 49, No. 3, 2006 To attain appropriate coefficients of L 0, L 1, L 2, one can find the inductances of any three different angles of rotor, such as θ = θ 0, θ = θ 1, θ = θ 2, and then from (1) it results in L(θ 0 ) 1 1 1 ( L(θ 1 ) = αnr ) L 0 1cos cos(αn r ) L 1 L(θ 2 ) 2 1 cos(αn r ) cos(2αn r ) L 2 L 0 = Ω L 1 (5) L 2 and thus L 0 L(θ 0 ) L 1 = Ω 1 L(θ 1 ) (6) L 2 L(θ 2 ) Once L 0, L 1, L 2 are obtained, the inductances (2) (4) of stator poles can be analytically expressed. As for the inductances L(θ 0 ), L(θ 1 )andl(θ 2 ), they can be attained by the FEM or some other analytical modeling methods. Although the FEM can provide more precise results than analytical modeling methods, it cannot parameterize the output performance directly, just like an analytical model. This paper adopts the permeance model to derive the inductances of the DSRM at three representative positions: aligned (θ 0 = 0), midway (θ 1 = α/2) and unaligned (θ 2 = α). As usual, the inductance is assumed only related to the angle of rotor and defined as L(θ) = N2 R(θ) = N2 P(θ) (7) where N is the number of turns in each phase, R(θ) isthe reluctance and P(θ) is the permeance. The total permeance is the sum of lumped permeances, such as parallel surfaces, non-parallel surfaces, etc. (7) The lumped permeances of the DSRM at three positions are constructed and shown in Appendix A in detail. By substituting the geometries in Table 2 into (10) (18), the inductances L(θ 0 ), L(θ 1 )andl(θ 2 ) and the coefficients L 0, L 1 and L 2 are achieved and listed in Table 3. Thus, from (2) (4), the three phase inductances L a (θ), L b (θ), L c (θ) can be obtained and will be used to calculate the reluctance torque of DSRM in section 5. 4. 2 Reluctance torque of the DSRM The reluctance torque of SRM is resulted from the tendency of moving the rotor poles in line with the energized stator poles, i.e. to maximize the inductance of the JSME International Journal

885 Fig. 3 Supplied voltage and current with advancing turn-on and turn-off angles excited coil. Generally, the reluctance torque T canbe express as (1) T = 1 dl(θ) i2 (8) 2 dθ where i is the exciting current. For this DSRM, the overall torque is obtained by T fld = 1 2 P S k=1 i 2 dl k (θ) k dθ where T fld is the overall torque, P S is the number of stator poles, i k is the energizing current of the kth stator pole, θ is the mechanical angle of the rotor and L k (θ) is the inductance of the kth stator pole obtained from (2) (4). 4. 3 Control strategies From (9), the simplest way to drive the DSRM is to energize the kth stator pole when the inductance is increasing,andtoswitchoff the supplied voltage when the inductance is decreasing. However, such simplest way cannot provide the desired performance with larger output power and torque. Thus, in order to achieve larger output power and torque, two parameters, the turn-on angle θ on and the turn-off angle θ off, are needed to be determined, as shown in Fig. 3. At high speed, the torque can be increased by advancing θ on, i.e. larger current can be established when the inductance is increasing. Besides, by advancing θ off before the inductance starts to decrease, negative torque can be prevented and the efficiency can be improved. By adjusting these two parameters according to different operation speeds, the torque can be enlarged, overall efficiency can be improved, and the constant power range can be extended (2). 5. HIL Simulator and Simulation Results The HIL simulator is built based on Matlab/Simulink with two computers, one used as the on-line controller and the other for simulating the scooter dynamics. The (9) control command and the simulation data transferred between these two computers are communicated through I/O boards (MRC-6810 and NI-6310) in real-time. Figure 4 shows the block diagram. The Model-PC consists of the energy storage system, power converter, the model of DSRM, transmission and the scooter dynamics. Note that the scooter dynamics includes rolling friction, aero drag force, etc. Inside the Model-PC, power in the energy storage system is fed into the DSRM by modulating the power converter. Then, the output torque of DSRM is computed to drive the scooter through the transmission. As for the Control-PC, it is used to log the feedback data to compute the suitable switching signals, and show the simulation results. When models become more complicated, three or more computers could be used simultaneously to precisely simulate the individual parts of this system. 5. 1 Output performance of the DSRM For a hybrid scooter, the specifications of electric motors are determined from the drive cycle. In this paper, it is required to track the ECE40 drive cycle below 25 km/h for the hybrid scooter driven by the DSRM. Therefore, the DSRM is designed to satisfy the following specifications: the constant torque region ranging from standstill to 6 500 rpm and the constant power region ranging from 6 500 rpm to 10 000 rpm. Within the constant torque region, the hysteresis-type current control method is adopted with an upper limit and a lower limit (1). Simulation results of the current waveform and torque of one single phase are compared at different speeds: 500, 1500, 4500 and 6 500 rpm with constant turn-on angle (θ on = 0 )andturnoff angle (θ off = 2 ) in Fig. 5. Obviously, when speed increases, the positive torque will be reduced, and negative torque marked by gray rectangle will be induced until the energized current does fade out. Thus, constant θ on and θ off may not lead to the desired performance unless they are variable in different operation conditions. Moreover, as shown in Fig. 5 (b), upper limit and lower limit of the hysteresis band are 16 A and 15 A, and with constant input voltage, the switching frequency decreases as the phase inductance increases. Since the output performances are affected by θ on and θ off significantly, Fig. 6 shows the output power and torque with θ off fixed and θ on varying from 0 to 10, and Fig. 7 shows the results with θ on fixed and θ off varying from 0 to 10. Evidently, advancing θ on and θ off will lead to greater output power and larger torque at high speed. By suitably adjusting θ on and θ off, the DSRM will provide better output performance. From above, the hysteresis-type current control strategy is employed within the constant torque region. In order to obtain maximal torque and prevent the negative torque, θ on and θ off must to be adjusted as well. However, within the constant power region, θ on and θ off are the only control parameters. Thus, the control objective is to find JSME International Journal Series C, Vol. 49, No. 3, 2006

886 Fig. 4 The block diagram of HIL simulator suitable θ on and θ off to extent the constant power region. Repeating the simulation procedure with θ on and θ off varied from 0 to 26, the suitable values of θ on and θ off can be determined at speed from standstill to 10 000 rpm, as shown in Fig. 8. By taking values of θ on and θ off into the switching strategy, the desired average torque and output power are achieved and shown in Figs. 9 and 10. From standstill, the average output torque in Fig. 9 keeps around 0.625 Nm due to the upper current limit and the average output power in Fig. 10 increases as the speed increases. Once the average output power reaches 420 W, the speed is around 6 500 rpm. Above 6 500 rpm, the output torque in Fig. 9 decreases since the DSRM operates in the constant power region, and the average output power keeps constant around 420 W. 5. 2 Drive cycle test For a hybrid scooter running in the urban area, the motor is used as the main power source when its speed is below the range from 17.5 km/h to 20 km/h, while running beyond this range, the gasoline engine is used instead. Figures 11 and 12 show the HIL simulation results of the hybrid scooter driven by DSRM only. In Fig. 11, it is clear that when the speed command is below 25 km/h, the tracking performance follows the ECE40 drive cycle efficiently. While for the speed over 25 km/h, Series C, Vol. 49, No. 3, 2006 the DSRM cannot supply enough power for the scooter to follow the ECE40 drive cycle. In Fig. 12, the total power demanded in ECE40 drive cycle shows that for the speed over 25 km/h or larger acceleration, the gasoline engine should be adopted as the main power source. Moreover, if the scooter can follow ECE40 drive cycle well, the negative power in deceleration region can be further used as regenerating power (3). Evidently, Figs. 11 and 12 demonstrate that the DSRM is a useful auxiliary power source for the hybrid scooter. 6. Conclusion In this paper, the DSRM is designed and used as an auxiliary power source for the hybrid scooter. With the axial flux distribution, it provides higher torque/volume ratio and higher power density in this application. The analytical model of the DSRM derived from permeance model can provide the dynamics of the torque and current. Furthermore, the analytical model can be used to optimize the geometries of the DSRM or modify them to fit some specifications. Till the output of the DSRM close to the desired performance, the FEM analysis is then processed for detailed modifications. Based on the HIL simulation environment, the components of the hybrid scooter can be combined together. Consequently, the feasibility of the JSME International Journal

887 (a) (b) Fig. 5 (c) (d) Phase current and torque: (a) 500 rpm (b) 1 500 rpm (c) 4 500 rpm (d) 6 500 rpm Fig. 6 (a) (b) (a) Output power and (b) output torque, with variable turn-on angle θ on Fig. 7 (a) (b) (a) Output power and (b) output torque, with variable turn-off angle θ off JSME International Journal Series C, Vol. 49, No. 3, 2006

888 Fig. 8 The turn-on and turn-off angles of the DSRM at different speeds Fig. 11 ECE40 speed tracking performance (driven by DSRM only) Fig. 9 The average output torque of the DSRM versus speed Fig. 12 Total power demanded in ECE40 drive cycle and power provided by DSRM hybrid scooters with simple modifications. More importantly, the hybrid scooter can improve the fuel economy, reduce the air pollution in the urban areas and also extend the mileage limitation of the existing electric scooters. Acknowledgements The work in this paper was supported by a grant provided by National Science Council, Taiwan, R.O.C. (NSC 93-2213-E-009-133-). Appendix A Fig. 10 The average output power of the DSRM versus speed DSRM can be verified rapidly and repeatedly. Then, the prototype of the DSRM will be obtained to match the desired performance closely. Finally, by utilizing the developed DSRM, one can reform existing engine scooters into Series C, Vol. 49, No. 3, 2006 The total permeance of the DSRM is the sum of the lumped permeances. The total permeance is denoted as P total, and the lumped permeances are denoted as P 1, P 2, etc. as shown in the Figs. 13 15. Moreover, R and r denote the lumped permeances belong to outer stator pole and the inner stator pole respectively. At aligned position (θ = 0), the total permeance can be obtained by JSME International Journal

889 Fig. 13 Lumped permeances at aligned position (θ = 0) Fig. 14 Lumped permeances at midway position (θ = α/2) Fig. 15 Lumped permeances at unaligned position (θ = α) 1 P total (0) = 2 2 + (10) P total r P total R P total r = P 1 r +2P 2 r + P 31 r + P 32 r (11) P total R = P 1 R +2P 2 R + P 31 R + P 32 R (12) At midway position (θ = α/2), the total permeance can be obtained by 1 P total (α/2) = 1 1 + (13) P total r P total R P total r = (P 1 r + P 31 r + P 32 r )/2+ P 1 r + P 2 r + P 4 r + P 51 r + P 52 r (14) P total R = (P 1 R + P 31 R + P 32 R )/2+ P 1 R + P 2 R + P 4 R + P 51 R + P 52 R (15) At unaligned position (θ = α), the total permeance can be obtained by 1 P total (α) = 1 1 + P total r P total R (16) P total r = P 1 r + P 2 r + P 4 r + P 51 r + P 52 r (17) P total R = P 1 R + P 2 R + P 4 R + P 51 R + P 52 R (18) References ( 1 ) Miller, T.J.E., Switched Reluctance Motors and Their Control, (1993), Magna Physics Publishing, Oxford and Clarendon Press. ( 2 ) Rahman, K.M., Fahimi, B., Suresh, G., Rajarathnam, A.V. and Ehsani, M., Advantages of Switched Reluctance Motor Applications to EV and HEV: Design and Control Issue, IEEE Trans. Industry Applications, Vol.36 (2000), pp.119 121. ( 3 ) Fahimi, B., Emadi, A. and Sepe, R.B., Jr., A Switched Reluctance Machine-Based Starter/AlternatorforMore Electric Cars, IEEE Trans. Energy Conversion, Vol.19, No.1 (2004), pp.116 124. ( 4 ) Profumo, F., Zhang, Z. and Tenconi, A., Axial Flux Machines Drives: A New Viable Solution for Electric Cars, IEEE Trans. Industry Application, Vol.44, No.1 (1997), pp.39 45. ( 5 ) Huang, S., Luo, J., Leonardi, F. and Lipo, T.A., A Comparison of Power Density for Axial Flux Machines Based on General Purpose Sizing Equations, IEEE Trans. Energy Conversion, Vol.14 (1999), pp.185 192. ( 6 ) Cavagnino, A., Lazzari, M., Profumo, F. and Tenconi, A., A Comparison between the Axial Flux and the Radial Flux Structures for PM Synchronous Motors, IEEE Trans. Industry Application, Vol.38 (2002), pp.1517 1524. ( 7 ) Roters, H.C., Electromagnetic Devices, (1941), John Wiley, New York. ( 8 ) Fahimi, B., Suresh, G., Mahdavi, J. and Ehsani, M., A New Approach to Model Switched Reluctance Motor Drive Application to Dynamic Performance Prediction, Control and Design, Power Electronics Specialists Conference, PESC 98 Record. 29th Annual IEEE, Vol.2 (1998), pp.2097 2102. JSME International Journal Series C, Vol. 49, No. 3, 2006