Influence of Constructive Parameters on the Cogging Torque in PMSMs

P. M. Garca1, J. A. Gemes2, V. Moreno1 and A. M. Iraolagoitia2

Department of Electrical Engineering University of the Basque Country E.U.P., Plaza Europa 1, 28018 Donostia-San Sebastin (Spain) Phone: +34 943 017 234, fax number: +34 943 017 130, e-mail: pedromaria.garcia@ehu.es, vicente.moreno@ehu.es E.U.I.T.I., Plaza de la Casilla 3, 48012 Bilbao (Spain) Phone: +34 946 014 363, fax number: +34 946 014 300, e-mail: joseantonio.guemes@ehu.es, ana.iraolagoitia@ehu.es
2 1

Abstract. The influence of constructive parameters on the cogging torque of two permanent magnet synchronous motor with single and fractional slot winding is analyzed in this work. The number of stator slots is fixed and the parameters modified in both types of motor are: a) slot opening width, b) rotor pole radius, and c) magnets thickness. The simulation of the behaviour of the motors is done by the use of finite elements. The saturation of the core is considered in the calculation of magnetic field. The torque is calculated using the virtual work method. The results of both types are compared and discussed. Keywords. - Cogging torque, finite element, permanent magnet synchronous motors, virtual work. 1. Introduction
One of the main problems presented by permanent magnet synchronous motors (PMSMs) is the torque ripple being the cause of noise, vibration, speed variation and problems in control devices [1]. Minimizing this ripple is very important in the design process of these machines. One of the main causes of torque ripple is the cogging torque. Cogging torque is the result of interaction (magnetic attraction) between the magnetic flux from the magnets and stator geometry, producing a variable reluctance with angular position of the rotor. This torque (independent from the stator current), does not contribute to net torque of the machine and is manifested by the tendency of the rotor to align in a number of positions. Several researchers [1] - [4] have proposed various methods to reduce cogging torque (using a fractional number of slots per pole, skewing of magnets and / or stator slots, displacing and shaping the magnets, optimizing the magnet pole-arc-to-pole-pitch ratio, introducing auxiliary slots or teeth, etc. ..)

The use of a fractional number of slots per pole reduces the amplitude of cogging torque [5] with no use of techniques that, as it happens with skewing of the magnets / or the stator slots, result in an increase of difficulties in motor manufacturing. This reduction is achieved by increasing the number of torque cycles, since each slot is located in a different angular position with respect to the magnets [5]. The main purpose of this paper is to analyze the influence in cogging torque of: a) slot opening width, b) rotor pole radius, and c) magnets thickness, in two PMSMs. Both machines have same power and external dimensions, the only difference being rotor pole number (single and fractional slot winding). The finite element technique is used to simulate the machine behaviour. Core saturation is directly considered in the magnetic field calculation. The virtual work method is used to calculate the torque.

2. Development
A) Cogging torque Cogging torque can be expressed mathematically by the following equation [1], [2]:

1 2 d Tcog = g 2 d
where

(1)

is the magnetic flux density, is the air-gap

reluctance and

the rotor angular position.

Since reluctance varies periodically, can also be expressed as a Fourier series in the form [1], [3], [5]:

Tcog =

Tmk sen( mk )
k =1

(2)

where, Tmk is a coefficient which indicates the amplitude of each component of the series, m is the least common multiple between the number of stator slots and pole number of the machine and k is an integer. It can be proved that the number of cogging torque cycles per revolution of the rotor is equal to m. Given the cyclical nature of cogging torque is not necessary to study it for a full rotation of rotor, reducing the study to smaller angles, for example, the stator slot pitch angle. The number of periods of cogging torque during rotation of one slot pitch (N), can be determined using the following expression [2], [6]:

TABLE I. Details of the analyzed motors Connection Regulator input voltage (V) Power rating (Referred to 16 Hz) Current rating (A) Number of poles Number of stator slots Rotor outside diameter (mm) Stator inner diameter (mm) Stator outer diameter (mm) Permanent magnet material Width of magnet (mm) Length of magnet (mm) Coercive Force (kA/m) Remanent flux density (T) Star 400 6 kW 16,6 12 / 10 36 158 160 240 NdFeB 6 32/39 920 1,16

N=

2p GCD (2 p, Q )

(3)

where GCD is the maximum common divisor between number of poles (2p) and slot number (Q). A way to influence in the cogging torque of a machine is to find a relationship between Q and 2p producing a high number of cycles, this means a decrease in the peak-topeak value of cogging torque. B) Virtual work This method is based on calculating the force executed by one component of a system in a certain direction, analyzing variation of the energy stored in the system when the component is moved in the given direction. Mathematically can be expressed as:

Table II shows number of stator slots (Q), number of poles, number of slots per pole and phase (q), number of cogging torque cycles per revolution of the rotor (m) and number of cycles per stator slot pitch (N) of the two investigated machines.
TABLE II. Number of cycles of cogging torque Q 36 36 2p 12 10 q 1 1,2 m 36 180 N 1 5

Fp =

W ( , p ) Wc ( i , p ) = p p = const i = const

(4)

where Fp is the total force exerted in p direction, W is the magnetic energy stored and Wc is the magnetic coenergy. In this way, the global torque developed by a rotating electric machine can be calculated from the magnetic coenergy, more specifically, from its partial derivative with respect to the angular displacement of the rotor [7].

From the values listed in Table II, we can expect a significant difference in the two motors behaviour in relation to cogging torque. In the first one (2p = 12 and Q = 36), with a integer number of slots per pole and phase, the number of cycles of cogging torque per slot pitch is 1, while the second motor (2p = 10 and Q = 36), with a fractional number of slots per pole and phase (1.2), has a number of cycles of cogging torque by slot pitch of 5. According to this, the second motor will give a lower cogging torque.

4. Analysis and results

The finite element technique is used for the computation of machine behaviour. A nonlinear field analysis is carried out for the magnetic flux density calculation in every nodes singular of the model. The finite element model used is two-dimensional. The use of parametrization techniques can deal with each dimension of machine's geometry as a variable, allowing faster determination of the finite element models, so that we can modify the parameter and determine, with each calculation process, the influence on torque of modified constructive parameter. Cogging torque is calculated by simulating the machine behaviour on no-load (without current in the stator windings) for different rotor positions. For each rotor position, model mesh is automatically recalculated. Fig 1 shows the magnetic flux density distribution of the two analyzed motors, for a fixed rotor position.

T=

Wc

(5)

where is the rotor angular displacement.

3. Analyzed motors
The analyzed machines in this study are of the embedded magnets type, both in the same power and external dimensions, the only difference being the number of rotor poles (12 and 10). Table I shows main design parameters of both analyzed motors.

A) Influence of slot opening width This section analyzes the influence of the slot opening width for values between 2 and 4 mm. Figure 3 shows, for the two motors analyzed in this paper, the cogging torque as a function of rotor position for different slot opening widths.
35 28

Color Shade Results

Quantity : |Flux density| Tesla Time (s.) : 15 Pos (deg): 15 Scale / Color 2,84258E-9 / 152,8144E-3 152,8144E-3 / 305,62884E-3 305,62884E-3 / 458,44322E-3 458,44322E-3 / 611,25761E-3 611,25761E-3 / 764,07206E-3 764,07206E-3 / 916,88645E-3 916,88645E-3 / 1,0697 1,0697 / 1,22252 1,22252 / 1,37533 1,37533 / 1,52814 1,52814 / 1,68096 1,68096 / 1,83377 1,83377 / 1,98659 1,98659 / 2,1394 2,1394 / 2,29222 2,29222 / 2,44503

2,0 mm 2,6 mm 3,0 mm 3,6 mm 4,0 mm

0 2 4 6 8 10

C o g ging to rq u e (N m )
a) 12-pole motor

21 14 7 0 -7

-14 -21 -28 -35

Rotor position (mechanical degrees)

a) 12-pole motor
0,12

2 mm
0,09
Quantity : |Flux density| Tesla ACH_GARGANTA : 3 Time (s.) : 0,1 Pos (deg): 99,999E-3 Scale / Color 7,84535E-9 / 152,45765E-3 152,45765E-3 / 304,91531E-3 304,91531E-3 / 457,37296E-3 457,37296E-3 / 609,83062E-3 609,83062E-3 / 762,28827E-3 762,28827E-3 / 914,74599E-3 914,74599E-3 / 1,0672 1,0672 / 1,21966 1,21966 / 1,37212 1,37212 / 1,52458 1,52458 / 1,67703 1,67703 / 1,82949 1,82949 / 1,98195 1,98195 / 2,13441 2,13441 / 2,28686 2,28686 / 2,43932

3 mm 4 mm

C o g g in g to rq u e (N m )

Color Shade Results

0,06 0,03 0,00 0 -0,03 -0,06 -0,09 -0,12 2 4 6 8 10

Rotor position (mechanical degrees)

b) 10-pole motor b) 10-pole motor Fig. 1. Magnetic flux density. Fig. 3. Cogging torque as a function of slot opening width, 70 mm pole radius and 6 mm magnets thickness.

The effect of the slot opening width, rotor pole radius and magnets thickness (see Fig 2) on cogging torque for both motors mentioned above is discussed in this section.

We can see that in the 12-pole machine there is a clear influence of slot opening on maximum torque, with a slot opening width of 2 mm the maximum value of cogging torque is reduced by 64% compared to the value corresponding to a slot opening width of 4 mm, this relationship is linear (see Fig 4). However, this is not so clear for the 10-pole machine, showing maximum cogging torque for the smaller opening width (2 mm). The reduction of maximum torque for the 10-pole motor and 3 mm slot opening width, compared to 12-pole motor and similar slot opening width is 99%.

Fig. 2. Analyzed parameters.

35

30

25

20

15

10 2,0 2,5 3,0 3,5 4,0

As it can be seen, in this case the variation of the maximum torque with rotor pole radius is similar in both machines. For the 12-pole machine, cogging torque magnitude is reduced by 35% for 65 mm radius compared to the cogging torque magnitude with 79 mm radius. The same comparison for the second motor, gives even greater differences, with an 85% reduction in maximum torque values (torque value is so small for pole radius between 65 and 70 mm, that calculation errors have a significant influence on the result).
Slot opening (mm)

Maximum torque (Nm)

Fig. 4. Maximum cogging torque as a function of slot opening width (12-pole machine).

The evolution of peak values of cogging torque for both machines, are shown in Fig. 6.

B) Influence of rotor pole radius This section examines the influence of pole radius (Fig. 2), for values between 79 and 65 mm at intervals of 1 mm, maintaining a constant rotor outside diameter of 158 mm. Figure 5 shows the cogging torque obtained for both machines, indicating a clear influence of polar radius on the maximum torque. The maximum value increases with the pole radius, may be because for these radius values, the air-gap width at the ends of the polar areas is lower and more constant than in smaller radius cases.
30

M axim un torque (N m )

29 27 25 23 21 19 17 15 65 67 69 71 73 75 77 79

Pole radius (mm)

a) 12-pole motor
0,3 0,3 0,2 0,2 0,1 0,1 0,0 65 67 69 71 73 75 77 79

65 mm
20

C ogging torque (N m )

67 mm 70 mm 73 mm 76 mm 79 mm

10

0
0 2 4 6 8 10

-10

Maximum torque (Nm)

-20

Pole radius (mm)

b) 10-pole motor

-30

Rotor position (mechanical degrees)

Fig 6. Maximum cogging torque as a function of polar radius.

a) 12-pole motor
0,3 0,2 0,1 0,0 0 2 4 6 8 10
65 70 75 79

C) Influence of magnet thickness This section examines the effect that magnets thickness have on cogging torque for values between 5 and 7 mm. Fig. 7 shows the results obtained for both machines. As it happened with the first parameters (slot opening width), evolution of maximum torque as a function of magnet thickness is different on both analyzed machines, clearly increasing in the 12-pole machine and decreasing in the 10-pole machine. The evolution of the maximum value of cogging torque for the 12-pole machine is again, as preceding paragraphs, almost linear (see Fig 8). Maximum torque value is reduced 14% when reducing magnet thickness from 7 to 5 mm.

Cogging torque (Nm)

-0,1 -0,2 -0,3

Rotor position (mechanical degrees)

b) 10-pole motor Fig. 5. Cogging torque as a function of polar radius, slot opening width of 3 mm and 6 mm magnet.

24 18

5,0 mm 6,0 mm 7,0 mm

Cogging torque (Nm)

12 6 0 0 -6 -12 -18 -24 2 4 6

10

Performance of various constructive parameters as a method to reduce cogging torque is much higher on machines with a integer number of slots per pole, where evolution of cogging torque with the studied parameter is more predictable. The studies showed important differences between the values of maximum torque calculated. In this motor, cogging torque decreases significantly with opening slot width, polar radius and magnets thickness. These claims are invalid for the case of the fractional slot winding machine. The results show a less predictable relationship between the cogging torque and the studied parameters. Only when the polar radius is modified its evolution is similar to that of the 12-pole machine.

Rotor position (mechanical degrees)

a) 12-pole motor

6. References
0,10

5,0 mm
0,08

6,0 mm 7,0 mm

0,05 0,03 0,00 0 2 4 6

10

-0,03 -0,05 -0,08 -0,10

Rotor position (mechanical degrees)

b) 10-pole motor

Fig 7. Cogging torque as a function of magnet thickness, 3 mm slot opening and 70 mm pole radius.
23

22

21

20

19 5 5,5 6 6,5 7

magnet thickness (mm)

Fig 8. Maximum cogging torque as a function of the magnet thickness (12-pole machine).

5. Conclusions
This paper analyzes, using the finite element method, the influence of slot opening width, pole radius and magnet thickness on cogging torque of two PMSMs. Torque values calculated on fractional slot winding machine are much smaller than those calculated on the single slot winding machine, justifying the use of fractional slots-per-pole number as an effective method for reducing cogging torque in this type of machines.

[1] L. Dosiek, P. Pillay, Cogging torque reduction in permanent magnet machines, IEEE Trans. Industry Application, vol. 43, no. 6, pp. 1656 1571, 2007. [2] D. A. Gonzalez, J. A. Tapia, A. Letelier, Desing Consideration to Reduce Cogging Torque in Axial Flux Permanent-Magnet Machines, IEEE Trans. Magnetics, vol. 43, no. 8, pp. 3435 3440, 2007. [3] Z. Q. Zhu, Influence of Design Parameters on Cogging Torque in Permanent Magnet Machines. IEEE Trans. Energy Conversion, vol. 15, no. 4, pp 407-412, 2000. [4] R. Islam, I. Husain, A. Fardoun, and K. McLaughlin, Permanent-magnet synchronous motor magnet designs with skewing for torque ripple and cogging torque reduction, IEEE Trans. Industry Applications, vol. 45, n. 1, pp. 152-160, 2009. [5] Z. Q. Zhu, S. Ruangsinchaiwanich, and D. Howe, Synthesis of cogging-torque waveform form analysis of a single stator slot, IEEE Trans. Industry Applications, vol. 42, no. 3, pp. 650-657, 2006. [6] N. Bianchi, S. Bolognani. Reducing Torque Ripple in PM Synchronous Motors by Pole-Shifting. ICEM 2000, pp 1222-1226. [7] T. Li, G. Slemon. Reduction of Cogging Torque in Permanent Magnet Motors. IEEE Trans. Magnetics, vol. 24, n. 6, pp 2901-2903, 1988. [8] M. S. Islam, S. Mir, T. Sebastian Issues in Reducing the Cogging Torque of Mass-Produced Permanent-Magnet Brushless DC Motors. IEEE Trans. Industry Applications, vol. 40, n. 3, pp 813820, 2004. [9] Z. Q. Zhu, S. Ruangsinchaiwanich, Y. Chen, D. Howe, Evaluation of Superposition Technique for Calculating Cogging Torque in Permanent-Magnet Brushless Machines, IEEE Trans. Magnetics, vol. 42, no. 5, pp. 1597-1603, 2006.

Maximum torque (Nm)

C ogging torque (N m )