Optimal Design and Experimental Verification of a Line-Start Permanent Magnet Synchronous Motor Guang Yang, Jun Ma, Jian-Xin Shen*, and Yu Wang College of Electrical Engineering, Zhejiang University, Hangzhou, 310027, China Email:
[email protected]
Abstract- This paper aims at optimal analysis and design of a three-phase line-start PMSM with simple structure, low cost and good performance. A prototype with 4 magnet poles is designed and manufactured. Simulation and experimental results are approximately the same and the prototype essentially satisfies the design request. Moreover, measures are taken to further optimize the design.
I.
The material of stator lamination is DR510-50, the parameters is shown in Table-III. TABLE I STATOR LAMINATION AND ARMARTURE WINDING DATA OF Y90S-4 INDUCTION MOTOR Item
INTRODUCTION
Stator Core
Line-start permanent magnet synchronous motor (LS-PMSM) is a kind of high efficiency synchronous motor. A squirrel cage is provided in the LS-PMSM to accelerate the rotor from standstill, whilst the magnets make the motor run at the synchronous velocity. Compared to the traditional three-phase induction motors, the LS-PMSM has typical advantages such as higher efficiency and power factor within a wide load range. This kind of motor could be a more efficient solution for general use applications than the induction motor. The LS-PMSM has been widely studied since it was first presented. A lot of experiences have been accumulated, and the machine has been applied to some particular industrial occasions. However, due to the complexity of the magnetic circuit and the rotor structure of the LS-PMSM, the optimal design method need further study. This paper describes a 4-pole prototype with a brief presentation of the design principle and procedure, and then presents the results of the steady-state and transient performance of the prototype. The experimental results are compared with the FEA results, and the probable reasons which could produce error are also analyzed. Moreover, the performance of the prototype LS-PMSM is experimentally compared with a Y90S-A induction motor, exhibiting significant improvement. II.
Armature Winding
Dimension 24 80mm 130mm 90mm 0.71mm Single-layer Y connection 81 12
No. of slots Inner diameter Outer diameter Length Wire diameter Winding mode Phase connection Coil turns per slot No. of coils
TABLE II STATOR NAME PLATE DATA OF Y90S-4 INDUCTION MOTOR Model:Y90S-4 2.7A IP44
1.1kW 1400r/min 20kg
380V 50Hz
PROTOTYPE DESCRIPTION AND DESIGN
A. Stator Structure The stator of the designed LS-PMSM uses the same lamination core and windings as 380V 4-pole 1.1KW Y-series induction motor which is denoted as Y90S-4, Fig.1. This kind of induction motor uses rotor slots skewed. To simplify the structure, skewing magnets is not used in the designed LS-PMSM. The data of stator lamination and armature winding are given in Table-I. The name plate data of stator are given in Table-II. The structure and winding mode are shown in Fig.1.
Fig.1. Stator lamination and winding distribution TABLE III MAIN PARAMETERS OF DR510-50 LAMINATION Grade DR510-50
Max. Iron Loss at 50Hz 20℃ 5.1W/kg
Min. B 1.64T
Min. Packing Factor 95%
Theoretical Density 7.55~7.75g/cm3
B. Rotor Design Rotor design is crucial for a LS-PMSM. After determining the dimension of air gap and rotor shaft, the residual space is
3232
employed for emplacing squirrel cage and magnets. These two parts have different functions under different states, thus, they should be designed respectively before integrated designing. Above all, using the simple 4-pole radially magnetized geometry is determined, and the inexpensive DR600-50 lamination is selected as there is almost no iron loss in the rotor core under steady-state condition. The squirrel cage is key point. The design of the squirrel cage mainly contains the dimension design and the number selection of rotor slot. It can make reference to the design principle of induction motor. There are many rotor slot types. In a LS-PMSM the squirrel cage is only functional when motor is starting up. When the motor is running at steady-state, the squirrel cage bar doesn’t cut the magnetic field, so it won’t generate additional rotor loss. Therefore, the design of the squirrel cage mainly focuses on good starting ability. For the purpose of satisfying different requirements under different situations, we have to make a compromise on the cage resistance: While cage resistance is relatively high, the motor will produce high starting-torque even under low starting-current. However, excessively high cage resistance will make it difficult to synchronize the motor speed; While cage resistance is relatively low, the synchronizing capability is better, but the starting-current will be higher. Pyriform slots with flat bottom are used on the rotor core for the purpose of reducing leakage coefficient; On the other hand, the average resultant torque/speed curve should have the steepest possible gradient near synchronous speed in order to get better synchronizing capability. This requires, among other things, a low cage resistance [1], therefore, too shallow or too narrow slot will be improper for the rotor. There are many restrictions for selecting the rotor slot’s number of induction motor. For example, in order to reduce the stray loss at the steady-state operation the near slot-combination is often adopted. However, being different form that of the induction motor, the squirrel cage of the LS-PMSM doesn’t function at the steady-state operation, the far slot-combination can be adopted, considering the influence of cage resistance and slot number on starting-ability of the motor, it is decided to be 36 slots in rotor, namely, 9 slots per one pole, the final designed squirrel cage of the rotor is shown in Fig.2. C. Magnet Design For the purpose of reducing the space which is occupied by magnets and making sure that the power density of the motor is high enough, sintered-NdBFe magnets are employed. The grade is N35SH. Detail parameters are given in Table-IV. The volume of the magnets mainly relies on the following factors: 1) Maximal magnetic energy product of magnets; 2) Operating Point and the torque current ratio which the motor requires; 3) Magnitude of back electromotive force; 4) Leakage coefficient; 5) Magnetic saturation;
The rated power of the motor is PN=1.1kW, and the rated frequency is f=50Hz. According to the magnetic energy product curve of N35SH and the expected magnetic energy utilization ratio, volume of the required magnets could be estimated. However, the impact of demagnetization should be considered, thus, the practical volume is a bit larger than the estimated one. The dimension of the rotor lamination length is designed to be 91mm, slightly larger than the stator core. Each pole consists of two pieces of 45.5mm-long magnets. The performance of the LS-PMSM varies with the magnet embedding depth. Therefore, it is essential to optimize the magnet embedding depth in order to improve the motor performance and reduce the magnet volume and cost. The researches about the influence of the magnet embedding depth have been reported in [2]. The influence has been inspected with finite element method while other geometry parameters keep constant, and some useful conclusions have been obtained. These will not be cited in this paper. The magnet embedding depth is determined as 23mm at last, which will bring relatively better performance of the motor. The final designed rotor geometry is shown in Fig.3. D. Air Gap Design Generally speaking, due to the strong magnetism in the PM motor, the air gap of a LS-PMSM is a little longer than that of a similar rating induction motor. Besides, proper increase of the air gap can reduce the vibration and noise, as well as the stray loss. Within a certain range, when the length of air gap increases by 0.1mm, the efficiency will increase by 1% [3], besides the increase of the air gap can also restrict the field saturation. However, if the air gap is too long, the rotor leakage will increase and will result in a decrease of loading capability and magnet utility efficiency.
Fig.2. Design of the rotor squirrel cage TABLE IV N35SH SINTERED-NDBFE MAGNETS DETAIL PARAMETERS Remanence
3233
1.18T
Coercive Force 876kA/m
Magnetic Energy Product 263~295kJ/m3
Working Temperature ≤150℃
Electric Conductivity 625kS/m
300 200 Phase Back EMF/V
Considering the empirical value of air gap length of the LS-PMSM which is usually 0.1~0.2mm more than that of the similar rating induction motor, we do research under the condition that the length of air gap varies from 0.35mm to 0.5mm as the air gap length of the Y90S-4 induction motor is 0.3mm. The research mainly includes the simulation and analysis of the influence of air gap length on the open-circuit back electromotive force (back EMF). As shown in Fig.4, while the air gap length varies form 0.35mm to 0.5mm, there is just a little influence on the back EMF. This is because the field generated by the magnets has already approached to saturation, in other words, under the restriction of rotor structure dimension, it will not obviously increase the back EMF only by increasing the volume of magnets or decreasing the air gap length. We have to increase the number of coil-turns in order to increase the back EMF. A relatively long air gap is chosen at last, which is 0.5mm. A long air gap is also helpful for assembly. However, it could deteriorate the starting performance. Thus, special attention should be paid. The simulated phase back EMF waveforms of the prototype are shown in Fig.5.
100 0 0.2
0.21
0.22
0.23
0.24
0.25
-100 -200 -300
time/ms
Fig.5. Simulated waveforms of phase back EMF when air gap length being 0.5mm
III.
TEST RESULTS OF LS-PMSM
A. No-Load Starting When the prototype is starting without load, the armature current waveforms of Phase A and Phase B are shown as Fig.6. The variety of armature current in starting process can be approximately divided into 4 stages: Stage I: Accelerating period. The rotor accelerates from standstill. Due to the low speed, namely slip is close to 1, the armature current is high, and the electromagnetic torque is also high, thus the rotor runs with high acceleration. Stage II: Approaching synchronization period. The speed rises continuously, and the slip is close to 0, but acceleration becomes lower and corresponding armature current decreases significantly; Stage III: Pulling into synchronization period. The motor goes into the damped oscillation procedure; Stage IV: Synchronous operation period. The waveforms of armature current tend towards stabilization after several fluctuations. During steady-state operation the armature current is inerratic sine-wave. 20
158
160
155
154
Phase A Phase B
15 Armature Current/A
Back Electromotive Force/V
Fig.3. Final designed rotor lamination of the Line-Start PMSM
154
120 80
Ⅰ
10
Ⅱ
Ⅲ
Ⅳ
5 0 -5
0
0.05
0.1
0.15
0.2
-10 -15
40
-20 0 0.3
0.35
0.4 0.45 Air Gap Length/mm
0.5
Fig.4. Back EMF RMS value curve versus air gap length
0.55
time/ms Fig.6. No-load staring current waveforms of prototype
B. Steady-State Operation Steady-state performance test of the prototype mainly includes the armature current, efficiency and power factor under load conditions. The measured armature current waveform and
3234
the simulated waveform at rated load are compared in Fig.7. The two curves agree well, verifying the simulation feasibility. Satisfactory agreement is also noticed under other load conditions, as shown in Fig.8. 6
By Experiment By Simulaton
2 0 0
0.005
0.01
0.015
0.02
-2 -4 -6
time/ms
Fig.7. Comparison of measured and simulated armature current waveforms at rated load
2.5 2 1.5
OPTIMIZATION OF LS-PMSM
Generally speaking, there are two methods to increase the PM-excited flux-linkage in the windings. One is to enhance the air-gap field by increasing the magnet size or grade. The other is to increase the number of turns of stator armature winding. In the prototype LS-PMSM there have been sufficient magnets, and increasing magnet size will hardly enhance the air gap field. Further, the grade of N35SH magnet has been sufficiently high. Therefore, the second method of increasing the number of coil-turns is employed. On account of restriction of slot-filling factor, we select thinner wire to increase coil-turns without changing stator structure. Thinner wire will make armature resistance of stator increase. However, it is predicated that the armature current will decrease significantly if the LS-PMSM is properly optimized. Finally, Φ0.62 copper wire is used to replace the original, and the number of coil-turns increases from 81 turns per slot to 90 turns per slot, remaining the slot-filling factor unchanged. The simulated back-EMF waveform of the optimized prototype is shown in Fig.11. Its RMS value increases from 154V in the original prototype to 167V in the new prototype at the rated speed. 3
1 By Simulation By Experiment
0.5 0 0
1
2
3 4 5 Load Torque/N•m
6
7
8
Fig.8. Comparison of measured and simulated armature current RMS value at different loads
From Fig.8 it is seen that the armature current tends to decrease and then increase when the motor is loaded up. This is because the PM-excited flux linkage in the winding is not sufficient. And, this will cause low power factor and high copper loss under no-load or light-load condition. Therefore, it is essential to increase the PM-excited flux-linkage. C. Comparison with Y90S-4 Induction Motor The steady-state performance of the same power Y-series induction motor, Y90S-4, is tested in order to compare with the prototype. Fig.9 shows the comparison of armature current RMS value at different loads between the LS-PMSM prototype and Y90S-4 induction motor. It is found that the armature current of the prototype is lower than that of the Y90S-4 within the load toque range from 4N·m to 7N·m, however, it is not the case when the load torque is lower than 4N·m. The efficiency comparison between the prototype LS-PMSM and Y90S-4 induction motor is shown in Fig.10. It is found that the prototype has a higher efficiency than the Y90S-4 within the range of load torque from 4N·m to 7N·m, but exhibits a lower efficiency within the 0-4N·m load torque range. The reason has been given in the preceding subsection. Therefore, the
3235
Armature Current RMS Value/A
Armature Current/A
3
IV.
2.5 2 1.5 1
LS-PMSM Before Optimization Y90S-4 Induction Motor
0.5 0 0
1
2
3 4 5 Load Torque/N•m
6
7
8
Fig.9. Comparison of armature current RMS value between prototype LS-PMSM and Y90S-4 induction motor
100 80 Efficiency/%
Current Waveform/A
4
LS-PMSM needs further optimization, basically by increasing the PM-excited flux-linkage in the windings.
60 40 LS-PMSM Before Optimization Y90S-4 Induction Motor
20 0 0
1
2
3 4 5 Load Torque/N•m
6
7
8
Fig.10. Comparison of efficiency between prototype LS-PMSM and Y90S-4 induction motor
90
200
80 Efficiency/%
Phase Back EMF/V
300
100 0 0.1
0.11
0.12
0.13
0.14
0.15
-100
70 60 50
LS-PMSM Before Optimization Y90S-4 Induction Motor LS-PMSM After Optimization
40 -200
30 0
-300
1
V.
CONCLUSION
[2]
Armature Current RMS Value/A
[3]
REFERENCE
T.J.E. Miller, “Synchronization of Line-Start Permanent Magnet AC Motors”, IEEE Transactions on Power Apparatus and Systems, Vol.PAS-103, No.7, pp. 1822-1828, 1984. J. Ma, S. Z. Dong, J. X. Shen, “Influence of Magnet Embedding Depth in Line-start Interior Permanent Magnet Synchronous Motor , ”Micro Motors, Vol.40, No.11, pp. 1-3, January 2007. R.Y.Tang, “Modern Permanent Magnet Machines—Theory and Design”, Beijing: Mechanism Industry Publishing Company, December 1997.
3 2.5 2 1.5 1 LS-PMSM Before Optimization Y90S-4 Induction Motor LS-PMSM After Optimization
0.5 0 0
1
2
3 4 5 Load Torque/N•m
6
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
6
7
7
8
LS-PMSM Before Optimization Y90S-4 Induction Motor LS-PMSM After Optimization 0
1
2
3 4 5 Load Torque/N•m
6
Fig.14. Comparison of power factor
Optimal design and experimental verification of a 4-pole 1.1kW LS-PMSM is presented. Its performance is significantly enhanced, whilst its additional cost can be compensated quickly. VI.
3 4 5 Load Torque/N•m
Fig.13. Comparison of Efficiency
Power Factor
Fig.11. Simulated back-EMF waveform of the optimized prototype
The optimized LS-PMSM is also tested and compared with the original LS-PMSM and the Y90S-4 induction motor, as shown in Figs.12-14. It is seen that the optimized LS-PMSM has the lowest armature current, highest efficiency and highest power factor in a wide range of load. Especially, when the load is more than half the rated value, the efficiency of the optimized LS-PMSM is about 7% higher than that of the Y-series induction motor, and the power factor is about 0.14 higher. Clearly, the performance improvement is rather significant. On the other hand, the material cost of the optimized LS-PMSM is 20% higher than that of the Y-series induction motor, which can be recovered after 550 hours full-load operation.
[1]
2
time/ms
8
Fig.12. Comparison of armature current RMS value
3236
7
8
