The Electrical Vehicle industry has developed rapidly in recent years, and the demand for new energy-driven motor systems has also increased. Permanent Magnet Synchronous Machine (PMSM) are high performance drive motors. In this paper, a new type of GO silicon steel material assembled stator core is proposed, compared with conventional Non Grain-Oriented (NGO) silicon steel stator, and the magnetic properties of the motor are analyzed. The method of assembling the stator core is given, in which the GO silicon steel sheets are stacked as the teeth of the stator core. GO silicon steel has low iron loss and large saturation magnetic density, which is more conducive to improving motor efficiency. The teeth of the stator core tend to have higher magnetic density and are more prone to sustain a higher iron loss. This research ensures the reliability of the test results through simulation analysis and physical test verification.

This paper presents numerical modeling and experimental measurement of core loss and thermodynamic characteristics of an iron core in PMSM at various frequencies among 50Hz, 100Hz, 400Hz and 1000Hz.

At first, differences of GO and NGO material will be described and metallographic experiment results are given. Articles and researches will be referred and several conclusions are verified in this section.

Secondly, a novel design strategy of PMSM will be described. The motor designed by authors is mounted on the vehicle and has 8 poles and 12 slots. The rated/peak power is 16/32kW. The numerical modeling in FEM is presented. The iron loss will be calculated, followed by measurement at different frequencies from 50Hz up to 400Hz. The initial measurement did not match the assumption and the reason will be described.

The third section shows improved method of experiment, which contained only one equivalent magnetic path. Power losses of NGO and its high-frequency iron core losses at 50Hz, 100Hz, 400Hz and 1000Hz are measured. It is clarified that the numerical model and experimental iron core losses in each experiment are showed to be in good agreement.

The numerical model for iron core losses and experimental results for single electrical steel sheet are especially useful to estimate power losses of electric machines running at higher frequency of 1000Hz 2000Hz. For different frequencies, the distribution of the thermodynamic model can be described by the iron core and the whole machine. Through the analysis and calculation, eddy-current loss is the main source that affects the thermodynamic characteristics of the motor. Thermos characteristic provides the direction for optimizing the cooling system of the motor. The final section presents our conclusion of the experiments and prototype’s performance.

Cold-rolled silicon steel sheets are divided into NGO silicon steel and GO silicon steel: GO silicon steel is further divided into general GO silicon steel and high magnetic induction GO silicon steel, namely HiB steel.1 NGO is mainly used to manufacture stator and rotor cores for electric motors.2 Generally, the magnetic flux density B and the core loss P value are used as the main indicators for measuring the magnetic properties of electrical steel products. At the same time, the amount of iron loss is also the basis for dividing the brand and naming of electrical steel products.3 Cold rolled GO silicon steel is excellent in magnetic properties because its Gaussian texture (ie,110<001>) is relatively large. This can simultaneously reduce iron loss and increase magnetic induction. Hysteresis loss decreases monotonically with the improvement of 110<001>’s orientation sharpness, which is represented by the increase of magnetic induction.4 Secondary recrystallization is the reason for the formation of Gaussian texture. The silicon content of GO does not change much, and the magnetic flux density increases only with the 110<001> deviation in grain orientation angle increase.5 The sustainable magnetic flux density is high, which can reduce the core excitation current, reduce the iron core loss and the loss caused by the wire resistance, and play a role of power saving.6 In the case of constant transformer or motor capacity, high magnetic induction can reduce the core volume and reduce the quality, save materials, reduce manufacturing costs, and facilitate in manufacturing, installation and transportation.

A comparison of metallographic structure between NGO and GO, after the first recrystallization and annealing, is shown in Fig. 1.

FIG. 1.

Comparison of Metallographic Between GO and NGO. (a) Metallographic structure of GO and (b) Metallographic structure of NGO.

FIG. 1.

Comparison of Metallographic Between GO and NGO. (a) Metallographic structure of GO and (b) Metallographic structure of NGO.

Close modal

Where it can be seen in the metallographic structure of NGO that the recrystallization has been already done, and the grain is larger. In the production process of GO, the initial recrystallization is followed by a high temperature and than a second-time cold-roll process. After the second cold-roll process, annealing is appled to form a favorable orientation-aligned Goss texture. The difference between the two material is essentially the difference in process and goal of the product.

The GO steel exhibits strong anisotropy in the rolling direction, and the grain structure is basically Gaussian texture and oriented in the rolling direction, so the GO steel in rolling direction has higher magnetic performance.

The NGO silicon steel sheet has the same magnetization performance in different direction, and the magnetic flux density saturation working point can reach about 1.6T. The uniformity of grain orientation in the GO steel sheet makes it excellent in electromagnetic properties in the rolling direction. The magnetic flux density saturation working point is generally above 1.9T, and the electromagnetic performance in the shear direction perpendicular to the rolling direction is higher. Therefore, GO silicon steel sheet’s magnetic flux density saturation operating point is generally around 1.5T, lower than NGO silicon sheet. The GO silicon steel has high magnetic permeability and low loss in the rolling direction. Generally speaking, GO silicon steel sheets are mainly used in transformers, large generators, etc. Transformers are used as stationary motors, during operation, the flux direction is only affected by the electric energizing signal, and there is no direction caused by mechanical motion.

On the other hand, in the operation of a rotating motor, the direction of the magnetic flux will change with the movement of the rotor. If GO is simply used in place of NGO in a rotating electrical machine, a sharp change in the direction of the magnetic flux at the joint of the stator yoke will result in deterioration of the performance of the motor. Since the saturation magnetic density of the stator tooth is larger than the stator yoke. This paper proposes a method of assembling by applying the GO to the teeth of the stator. During the operation, the magnetic flux direction of the stator tooth portion is parallel to the tooth direction, which can fully utilize the characteristics of GO in the rolling direction and improve the running performance of the motor. The numerical model of PMSM is shown in (Fig. 2).

FIG. 2.

Numerical Model of PMSM.

FIG. 2.

Numerical Model of PMSM.

Close modal

In magnetic performance measurement, firstly soft magnetic material AC test by ring specimens is applied directly to the iron core. The experiments has been done in the research centre of Tunkia Ltd., and in this method, magnetic field is formed by excitation current. The flux density of the magnetic field varies from 0.1T to 1.7T, the points in step of 0.1T are measured. Then, different frequency of excitation current from 50Hz to 400Hz in step of 50Hz are applied to GO iron core and conventional iron core. During the test, the system is usually an powered by an oscillating signal generator and power amplifier. A represents the rms value or peak current meter for measuring the magnetizing current, or the rms or peak voltmeter for mating with a static resistor; Hz stands for frequency meter; N1 represents the primary winding; N2 stands for secondary winding; OSC stands for oscilloscope; V1 represents the average voltmeter; V2 stands for rms voltmeter. And the calculation of magnetic field strength can be done by the formula below:

(1)

Where: H(t), represents the magnetic field strength corresponding to time t in ampere per meter (A/m); N1 represents the number of turns of the primary winding; U1(t) represents the voltage across the non-inductive precision resistor used to determine the magnetizing current for the corresponding time t, in volts (V);Lm represents the average annual magnetic circuit length of the sample in meters (m); R represents the resistance value of the non-inductive precision resistor used in series with the primary winding to determine the magnetizing current, in ohms (Ω).

The total loss Ps corresponds to the area of the hysteresis loop formed by B and H. The specific loss Ps per kilogram of the sample should be calculated according to Eq. (2):

(2)

Where: Ps represents the total loss of the sample in watts per kilogram (W/kg).

The eddy current loss of the sample can be measured by Eq. (3):

(3)

Where: Pe stands for eddy current loss; σ stands for the conductivity; γ stands for material density, t is the material thickness; Bm is the maximum working flux density.

The GO tooth core and the NGO conventional iron core are compared, and the changes of H and P with B value at different frequencies are measured respectively.

Take 50Hz low frequency and 400Hz high frequency measurement results for comparison. The test results are shown in the following figures (Fig. 3 and Fig. 4):

FIG. 3.

B-H Curves of 50Hz and 400Hz Comparison. (a) 50Hz B-H. (b) 400Hz B-P.

FIG. 3.

B-H Curves of 50Hz and 400Hz Comparison. (a) 50Hz B-H. (b) 400Hz B-P.

Close modal
FIG. 4.

Iron Loss Curves of 50Hz and 400Hz Comparison. (a) 50Hz B-P. (b) 400Hz B-P.

FIG. 4.

Iron Loss Curves of 50Hz and 400Hz Comparison. (a) 50Hz B-P. (b) 400Hz B-P.

Close modal

It can be seen that GO at low frequencies is inferior to that of NGO from both B-H and B-P. At high frequencies, the iron loss value of GO is reduced, which is the same as that of NGO. This phenomenon is analyzed as follows: 1) Degraded performance due to the influence of the gap at the joint of the yoke portion of the GO tooth portion; 2) Due to the difference in thickness of GO (B30P105 of thickness=0.3mm) and NGO (B20AHV1700 of thickness=0.2mm), the performance showed difference; 3) In the testing method, the direction of magnetic field is perpendicular to the rolling direction of GO sheet. Multiple magnetic paths resulting mismatching between the test and assumption. Therefore, it is necessary to establish a new model for experimental verification.

According to the analysis of the possible reason, which cause the failure of experiment result not matching the assumption. We concluded that reason 3) is the most possible reason. During the test, the magnetic path is a toroidal magnetic circuit, and is perpendicular to the teeth of the core. Therefore, when the tooth portion is made of GO material, the magnetic path direction is perpendicular to the rolling direction of GO material. This will result in a large difference between the measured value and the actual value. Hence, a new iron core model to measure the electromagnetic properties of GO material in the iron core is needed.

In this paper, we took one pole of a 12 slots 8 poles stator core as a model, and replace the middle tooth equivalent to both sides. The magnetic path of the rotor is replaced by a core block to form a closed circuit, as shown in the following figures (Fig. 5), where the arrow shows the magnetic path inside the block. This model uses the teeth made of GO as part of the loop to effectively avoid measurement errors caused by the direction issue. At the same time, the gap error caused by splicing process can be reduced, and the measurement result is more accurate. Based on this model, the following models were fabricated to test their magnetic properties and verify the difference in performance between GO and NGO steel in motor applications (as shown in Fig. 6).

FIG. 5.

Equivalent Model of Magnetic Circuit.

FIG. 5.

Equivalent Model of Magnetic Circuit.

Close modal
FIG. 6.

Experiment Grouping Situation.

FIG. 6.

Experiment Grouping Situation.

Close modal

The influence of the splicing process and the influence of the air gap at the core joint on the magnetic properties of the core are compared by group 1 and 2; magnetic properties of NGO steel sheet and GO steel sheet are compared by group 2 and 3; difference in magnetic properties caused by replacement of NGO stator iron core by combined stator core whose teeth is made of GO steel sheet are compared by group 1 and 4.

The result of B-H curves and B-P curves under different frequency are shown as followed (Fig. 7 and Fig. 8):

FIG. 7.

B-H-P Comparison of 4 groups. (a) B-H Comparison. (b) B-P Comparison.

FIG. 7.

B-H-P Comparison of 4 groups. (a) B-H Comparison. (b) B-P Comparison.

Close modal
FIG. 8.

High Frequency Core Loss Comparison. (a) 400Hz Iron Loss and (b) 1000Hz Iron Loss.

FIG. 8.

High Frequency Core Loss Comparison. (a) 400Hz Iron Loss and (b) 1000Hz Iron Loss.

Close modal

It can be seen from the that the B-H performance of group 1 and 4 is similar, and is better than group 2 and 3, indicating that the iron core will have a negative impact on the magnetic properties after the splicing process. Comparing group 2 and 3, group 2 is better than group 3 in both B-H performance and B-P performance, indicating that the magnetic properties of the iron core will be better after replacing the conventional NGO iron core with the GO steel for the tooth. Comparing group 1 and group 4, the magnetic properties of GO steels are slightly better than those of NGO steels, but their B-P characteristics are the same at high frequencies and high magnetic densities.

In order to verify the conclusion observed above, group 1 and 4 are separately observed, the iron loss of group 1 and 4 along with different frequency curve shows that iron loss in GO and NGO is not that different, the reason could be the magnetic path is perpendicular to the GO stator yoke, not in the rolling direction. The result of the figure is matched with our assumption. If GO fully replaces NGO stator core, the performance could be worse.

Combined with test 1 and modified test 2, the results show that: the GO silicon steel is indeed superior to that of the NGO steel in the rolling direction, but the magnetic properties in perpendicular of rolling direction are slightly inferior. The conclusion of the experiment matches with the assumptions made. For the iron loss characteristics, in the case of high frequency and high magnetic density, GO steel does not exhibit any advantage, and its iron loss is almost the same as that of NGO steel. However, the stator core made of GO silicon steel has a good performance in place of the conventional iron core made of NGO silicon steel.

In this section, thermal simulation of GO type motor and conventional NGO type motor has also be given. By the comparison between Fig. 9. Thermal simulation analysis of GO steel and NGO steel core is carried out and compared in peak conditions. The scheme adopts a water-cooled chassis and the cooling water flow rate is 8L/min. It can be seen from the figure that the maximum temperature of the iron core using the GO steel as the tooth portion is lower than that of the iron core using the NGO steel. The main reason is that due to the superior magnetic properties and lower iron loss of the GO steel, the current of is reduced, the copper loss is reduced, and the total loss is smaller than that of the NGO iron core, so that the temperature rise of the GO steel core is lower.

FIG. 9.

Comparison of Thermal Simulation Between GO Type and NGO Type Motor.

FIG. 9.

Comparison of Thermal Simulation Between GO Type and NGO Type Motor.

Close modal

This paper described the feasibility of replacing a conventional iron core made of NGO silicon steel with teeth made of GO silicon steel in a 12 slots, 8 poles motor. The teeth made of GO silicon steel are spliced to form a stator core, which increases the complexity of the process. However, GO silicon steel teeth in the motor is feasible in place of a conventional iron core made of NGO silicon steel in general speaking. In terms of performance, the iron core composed of GO silicon steel shows sufficient advantages. From the perspective of cost, the price of NGO silicon steel with low iron loss is similar to that of ordinary GO silicon steel, and the overall cost is controllable. In summary, the motor made of GO silicon steel in this solution has a certain mass production value.

1.
S.
Taguchi
,
T.
Yamamoto
, and
A.
Sakakura
, “
New grain-oriented silicon steel with high permeability “ORIENTCORE HI-B
,””
IEEE Transactions on Magnetics
10
,
123
127
(
1974
).
2.
K.
Matsumura
and
B.
Fukuda
, “
Recent developments of non-oriented electrical steel sheets
,”
IEEE Transactions on Magnetics
20
,
1533
1538
(
1984
).
3.
H.
Shimanaka
,
Y.
Ito
,
K.
Matsumara
, and
B.
Fukuda
, “
Recent development of non-oriented electrical steel sheets
,”
Journal of Magnetism and Magnetic Materials
26
,
57
64
(
1982
).
4.
Y.
Ushigami
,
M.
Mizokami
,
M.
Fujikura
,
T.
Kubota
,
H.
Fujii
, and
K.
Murakami
, “
Recent development of low-loss grain-oriented silicon steel
,”
Journal of Magnetism and Magnetic Materials
254
,
307
314
(
2003
).
5.
M.
Matsuo
, “
Texture control in the production of grain oriented silicon steels
,”
ISIJ International
29
,
809
827
(
1989
).
6.
A. W.
Lotfi
and
M. A.
Wilkowski
, “
Issues and advances in high-frequency magnetics for switching power supplies
,”
Proceedings of the IEEE
89
,
833
845
(
2001
).