Magnetic

Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

Available online at www.sciencedirect.

com
ScienceDirect

Energy Reports 8 (2022) 209–218


www.elsevier.com/locate/egyr

2021 The 2nd International Conference on Power Engineering (ICPE 2021), December 09-11,
2021, Nanning, Guangxi, China

High-speed permanent magnet machine multi-physics fields design


for turbine application
Yue Zhanga ,∗, Zhenning Qia , Huijun Wangb , He Zhangc
a Shandong University, Jinan, 250061, China
bBeihang University, Beijing, 100191, China
c University of Nottingham Ningbo China, Ningbo, 315100, China

Received 24 January 2022; accepted 14 February 2022


Available online 26 February 2022

Abstract
This paper presents the multi-physics fields research of a high-speed permanent magnet machine (HSPMM) for turbine
application. At first, the design considerations for the HSPMM are introduced and presented. Then, the machine electromagnetic
performances are comprehensively studied with machine stator slot numbers. Moreover, the influences from the PM
magnetization methods and segmentation number on PM eddy current loss reduction are further investigated. Thermal analysis
is also performed for the HSPMM with the cooling water fluid volume flow rate considered. In addition, rotor strength is
studied for the HSPMM with several design factors on machine rotor strength performances further studied and researched.
© 2022 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the scientific committee of the 2021 The 2nd International Conference on Power Engineering, ICPE, 2021.

Keywords: High-speed; Permanent magnet machine; Thermal analysis; Multi-physics fields

1. Introduction
High-speed permanent magnet machines (HSPMMs) are very promising and have been widely used in indus-
try [1–3]. The advantages of HSPMMs are categorized as compact size and high-efficiency [4]. However, the
compact size property also increases power loss density and heat dissipation difficulty, while over-heating challenges
the reliable operation of HSPMM [5,6]. Moreover, rotor strength analysis is necessary for HSPMM design, as
permanent magnet materials are low in mechanical strength and cannot sustain the large centrifugal force caused
by the high-speed rotation [7].
The HSPMM design with higher power density is highly desirable for this turbine application. At first, the
design considerations for the HSPMM are presented. Then the machine electromagnetic performances with power
losses are comprehensively researched to estimate the effects from stator slot number. Some papers have introduced
∗ Corresponding author.
E-mail address: yzhang35@sdu.edu.cn (Y. Zhang).

https://doi.org/10.1016/j.egyr.2022.02.176
2352-4847/© 2022 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http:
//creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the scientific committee of the 2021 The 2nd International Conference on Power Engineering, ICPE,
2021.
Y. Zhang, Z. Qi, H. Wang et al. Energy Reports 8 (2022) 209–218

the permanent magnet (PM) axial segmentation is helpful to reduce the PM eddy current loss [7,8]. This paper
extends the research to investigate the rotor PM eddy current loss variation with both PM segmentation number
and magnetization methods considered. The thermal performance for HSPMM is analyzed by Lumped method with
the cooling water fluid volume flow rate effects studied. The rotor strength is researched by Finite element method
(FEM) with both rotor sleeve thickness and interference fit effects on rotor stress studied.

2. HSPMM electromagnetic design

2.1. Machine design considerations

Surface-mounted PM rotor structure is increasingly used for high-speed applications, with consideration of its
simple structure and high mechanical strength properties. Fig. 1 presents the surface-mounted high-speed PM rotor
structure. PMs are brittle and mechanically weak. Therefore, rotor sleeve with interference fit is normally required
to protect the PMs against the large centrifugal stress induced by the high-speed rotation. Both non-magnetic metal
(such as stainless steel) and carbon fiber can act as sleeve, while stainless steel is used in this case due to its high
thermal conductivity characteristic. The number of poles for high-speed rotor is designed as 2 or 4 to reduce the
machine operating frequency. The rated power and speed for the designed HSPMM for turbine application is 30 kw
and 25 000 rpm, respectively. The main machine parameters, which are related to the machine stator inner diameter
Di and machine active length Le f , can be determined by the following equation:
6.1 P′
Di2 L e f = · (1)
α P K nm K d p ABδ n
where α P is the effective pole arc coefficient, Knm is the air gap flux factor, Kd p is the winding factor, A and Bδ
are the electrical loading and air gap flux density, while P′ is the apparent power and n is the machine speed. The
rotor outer diameter of HSPMM is determined by both the stress limits of the utilized materials and rotor rotation
circumferential speed, while machine stator outer diameter and active length are considered with installation space.
The pole pair number for high-speed machine is normally chosen from one or two to reduce machine operating
frequency. The operating frequency is low for the HSPMM with one pole pair, which is beneficial to reduce the
machine operating frequency and power losses. However, the weight of one pole pair machine is obviously increased
compared to the two pole pair counterpart, as the stator yoke should be designed as thick. As the HSPMM for turbine
application is desired to be a light weight one, thus, the pole pair number of the machine is set to be 2 in this case.

Fig. 1. Surface-mounted high-speed PM rotor.

The material and dimensions of PM poles have directly effects on the HSPMM performances. Considering the
difficulty in rotor thermal dissipation, the chosen PM material should be capable to operate steadily in a high working
temperature. In addition, it is also desirable for the chosen PM material with a high remanence and coercivity to
increase the machine power density. Hence, NdFeB material is a reasonable choice for the HSPMM in this case. The
PM thickness and span for the designed HSPMM are considered to increase the machine torque density. Machine
output power capability cannot be fully employed if the PM pole thickness is low. However, machine output torque
also declines if the PM poles are too thick, as it increases machine equivalent air gap length. Thus, for the HSPMM
in this case, the PM thickness is determined as 8 mm and the span is chosen as 1 to increase the machine power
density. The main parameters of the HSPMM for turbine application are listed in Table 1.
210
Y. Zhang, Z. Qi, H. Wang et al. Energy Reports 8 (2022) 209–218

Table 1. HSPMM main parameters.


Rated power (kW) 30 Rated speed (rpm) 25 000
Power supply frequency (Hz) 833.3 Stator outer diameter (mm) 116
Stator inner diameter (mm) 56 Air gap length (mm) 1
Pole pair number 2 Sleeve material Stainless steel
Sleeve thickness (mm) 1 PM material NdFeB
PM thickness (mm) 8 PM span 1
Winding layer number 2 Winding connection Y

2.2. Stator slot number

The 18 and 24 slots machine stator schemes are normally used for HSPMM stator. In this paper, these two stator
schemes are comprehensively researched and compared for the HSPMM in this application. Two HSPMMs with
18 and 24 stator slots are designed with the same stator outer and inner diameters, air gap length, sleeve and PM
thicknesses. The cross-sections of the two HSPMMs are presented in Fig. 2. The no-load performances for the
two machines are compared at first. The back EMFs for the two machines at rated speed are presented with FFT
analysis results in Fig. 3, while the cogging torque waveforms for the two machines are compared in Fig. 4.

Fig. 2. Cross-sections for the 18 and 24 slots HSPMM.

The voltage sin wave distortion rate (THD) is used to evaluate the harmonics in the back EMF waveform, which
is defined as:

E 32 + E 52 + E 72 + · · ·
THD = ∗ 100% (2)
E1
where E1 , E3 , E5 . . . ..represents the amplitude of fundamental, 3th order harmonic, 5th order harmonic and so on.
From the EMF analysis results of both HSPMMs, the THD value for the EMF of 18 slots HSPMM is 20.20%,
while the THD for 24 slots HSPMM is 15.67%. Thus, the harmonic components in the EMF of 24 slots HSPMM
are lower. In addition, the cogging torque of 24 slots HSPMM is also much lower than its 18 slots counterpart.
The performances for the two HSPMMs at rated load are also researched and compared. The electromagnetic
torque for PM machine Tem can be expressed as following:
( )
Tem = 1.5N P (ψ P M Iq + L d − L q Id Iq ) (3)
where N p and Ψ P M are the pole pair number and PM flux-linkage. While Ld , Lq , Id and Iq are the inductances and
currents in d- and q-axis, respectively. Considering the difference between Ld and Lq is negligible, as the saliency
ratio for surface-mounted PM rotor is small. Thus the HSPMM torque is the product of PM flux-linkage and q-axis
current. The two machines are powered by rated currents to obtain the similar rated output torque average value.
The output torque waveforms for the two HSPMMs at rated load are shown in Fig. 5. Clearly, the toque ripple of
the 18 slots HSPMM is higher than the 24 slots one, due to its higher harmonic components in the back-EMF.
The power losses of HSPMM mainly include copper loss, iron core loss, rotor eddy current loss and friction
loss. For HSPMM copper loss (Pcu ), it should be considered with both skin and proximity effects considered as the
winding current frequency is relatively high. As such loss is dependent on the magnetic field, thus, Finite Element
211
Y. Zhang, Z. Qi, H. Wang et al. Energy Reports 8 (2022) 209–218

Fig. 3. Back EMF waveforms with FFT analysis results for 18 and 24 slots HSPMMs.

Fig. 4. Cogging torque waveforms for 18 and 24 slots HSPMMs.

Fig. 5. Torque waveforms for 18 and 24 slots HSPMMs.

Method (FEM) is effective for HSPMM copper loss estimation correctly. Iron loss (Pir on ) is a major loss component
in the total power losses of HSPMM. It includes hysteresis loss (Ph ), eddy current loss (Pc ) and excess loss (Pe ),
which can be calculated as:
Pcor e = Ph + Pc + Pe = K h · f · Bmα + K c · f 2 · Bm2 + K e · f 1.5 · Bm1.5 (4)
where Kh , and α are the coefficients of hysteresis loss, Kc and Ke are the hysteresis loss coefficient and excess loss,
respectively; f is machine operating frequency and Bm is the flux density amplitude. Rotor eddy current loss causes
rotor temperature rise which may lead to HSPMM rotor overheating problem and PM demagnetization. Thus, it
needs to be estimated for HSPMM. Rotor eddy current loss is generated in machine sleeve and PM components,
212
Y. Zhang, Z. Qi, H. Wang et al. Energy Reports 8 (2022) 209–218

which can be calculated as:


|J |2

Peddy = dv (5)

where J is the eddy current density and σ is the material conductivity. Friction loss (P f riction ) is due to the friction
effect between rotor surface and air fluid when the rotor in high-speed rotation. It can be calculated as:
P f riction = kC f πρω3r 4l (6)
where k reflects the rotor roughness, C f is the air fluid friction coefficient, ρ is the air fluid density, r and l are the
rotor outer diameter and axial length, respectively. Then the efficiency of the HSPMM (η) can be calculated as:
Pout
η= (7)
Pout + Pcu + Pcor e + Peddy + P f riction
where Pout is HSPMM output power. The power loss performances for the two HSPMMs with different stator slots
at rated condition are compared in Table 2. It is found that the efficiency of the HSPMM with 24 stator slots is
slightly higher than the 18 stator slots one at rated condition. Moreover, it is noted that the rotor eddy current
loss of the 24 slots HSPMM is much lower than the 18 slots one, which is desirable to reduce the rotor operating
temperature and avoid PM demagnetization risk for the machine at high speed operation.

Table 2. Power loss comparison results for 18 and 24 slots HSPMM.


18 slots 24 slots
Iron loss (W) 1246.6 1192.3
Winding loss (W) 270.9 282.6
Sleeve loss (W) 84.3 26.8
PM loss (W) 11.0 1.3
Friction loss (W) 82.3 82.3
Total loss (W) 1695.1 1585.3
Output power (kW) 30 266.6 30 023.2
Efficiency 94.7% 95.0%

2.3. PM magnetization and segmentation for rotor eddy current loss reduction

This section performs a research in the magnetization method effects on HSPMM rotor eddy current loss. Three
magnetization methods, including radial, parallel and Halbach magnetizations (as shown in Fig. 6) are discussed.
The magnetic flux density distributions with flux lines for the HSPMMs with different magnetization methods at
no load and rated conditions are presented in Figs. 7 and 8, respectively. Permanent magnet axial segment method
is helpful to reduce the HSPMM eddy current loss in PMs, as the eddy current flowing path in magnets is cut off in
axial direction, due to the magnet segmentation [8]. For example, the magnet in four divisions is presented in Fig. 9.
This paper performs a comprehensive research on the axial magnets segment number on rotor eddy current loss
with different magnetization methods considered. The rotor PM eddy current loss for the HSPMMs with different
magnetization methods and axial segment numbers when the machine at rated condition are presented in Fig. 10.
It is learnt that when the magnet is not divided (namely, the PM segment number is one), the PM eddy current
loss of the HSPMM with PM radial magnetized is similar to the machine with PM parallel magnetized. While the
HSPMM PM eddy current loss is increased when the magnets are Halbach magnetized. Magnet segment is helpful
to reduce the HSPMM PM eddy current loss. However, the PM eddy current loss reduction effects by PM segment
number vary with different magnetization methods employed. The PM eddy current loss is greatly reduced with
the PM segments: when machine PMs are divided into four pieces in axial direction, the PM eddy current loss is
reduced by 75.5% and 74.5% when PMs are radial and parallel magnetized, respectively. However, such decline
trend is not obvious when PMs are Halbach magnetized: the PM eddy current loss is only reduced by 14.0% for
PMs in four pieces division. Thus, the magnet segmentation number needs to be increased for the HSPMM with
Halbach magnetized PMs.
213
Y. Zhang, Z. Qi, H. Wang et al. Energy Reports 8 (2022) 209–218

Fig. 6. PM magnetization methods.

Fig. 7. Magnetic flux density and lines distributions for the HSPMM at no load condition.

Fig. 8. Magnetic flux density and lines distributions for the HSPMM at rated condition.

Fig. 9. Magnets in four divisions.

214
Y. Zhang, Z. Qi, H. Wang et al. Energy Reports 8 (2022) 209–218

Fig. 10. PM eddy current loss with PM axial segment number.

3. HSPMM thermal analysis


Thermal analysis is necessary to be performed for HSPMM due to the thermal dissipation problem by the machine
compact size. Lumped method is utilized in this paper to analyze the thermal performance of the HSPMM in this
study. In order to constrain the HSPMM temperature rise, spiral water jacket around stator outer core for machine
cooling (the spiral water path is as shown in Fig. 11). The temperature calculation results for the HSPMM at
rated condition with the water fluid volume flow rate at 0.0005 m3 are presented in Fig. 12, while it is found the
maximum temperature of the machine locates at the rotor sleeve. The influences of cooling water fluid volume
flow rate on machine thermal performances are then researched. The variation of machine components maximum
temperatures with water volume flow rate for the HSPMM at rated working condition are studied and presented
in Fig. 13. It is found increasing the water flow rate is helpful to reduce the operating temperature of the machine
in a certain range. However, such temperature rise limit effect is limited after the flow rate above 0.0005 m3 , as
the temperatures in machine components are then almost steady. In addition, for the NdFeB PM material used for
HSPMM, the maximum permitted working temperature is around 160 ◦ C, while it is learnt from the results that
the HSPMM PM operating temperature is always below the permitted value within the whole range

Fig. 11. Spiral water path .


.

4. HSPMM rotor stress analysis


Normally, sleeve is utilized to mechanically protect rotor surface-mounted PM poles with interference fit for
HSPMM in high speed operation. For the non-magnetic alloy steel sleeve in this study, its thickness is 1 mm and
the interference fit is set as 0.1 mm. The mechanical properties of the non-magnetic alloy steel and PM material
for the HSPMM rotor are listed in Table 3, while the sleeve equivalent stress with PM radial stress and tangential
stress for the rotor operating at high temperature are presented in Fig. 14. The sleeve maximum equivalent stress
is 743.14 MPa for the rotor at rated speed, which is below the permitted value for the steel alloy (1100 MPa). The
rotor PM radial stress is a compressive one, while PM material is capable to sustain a large compressive stress
215
Y. Zhang, Z. Qi, H. Wang et al. Energy Reports 8 (2022) 209–218

Fig. 12. HSPMM temperature calculation results.

Fig. 13. Variation of machine components maximum temperatures with water fluid volume rate.

Table 3. Non-magnetic ally steel and PM mechanical properties.


Alloy steel Ndfeb
Density (kg/m3 ) 7850 7500
Thermal expansion coefficient (µm/m/◦ C) 12 10
Young’s modulus (GPa) 200 160
Poisson’s ratio 0.3 0.24

(800 MPa). However, it is noted that PM is also weak in tensile (75 MPa). For the HSPMM rotor PM tangential
stress, the maximum tensile stress is 12.3 MPa, which is below the PM permitted value. The rotor maximum sleeve
equivalent stress and PM stress variations with sleeve thicknesses and interferences are further presented in Fig. 15.
It is learnt the sleeve equivalent stress rises more obviously with the sleeve interference fit. In addition, increasing
sleeve thickness is helpful to reduce the rotor PM tensile stress in high-speed operation, especially with a large
interference fit.

5. Conclusions
This paper presents the multi-physics fields research work for a HSPMM for turbine application. At first, the
design considerations for the HSPMM are introduced. Then the influences of stator slot number on machine
performances are comprehensively researched. It is found the HSPMM with 24 stator slots provides lower torque
ripple and rotor eddy current loss than the 18 slots one. In addition, the HSPMM PM eddy current loss is investigated
216
Y. Zhang, Z. Qi, H. Wang et al. Energy Reports 8 (2022) 209–218

Fig. 14. Rotor stress distribution.

Fig. 15. Rotor maximum stress variation with sleeve thickness and interference fit.

with PM axial segment number and magnetization methods. It is found for the radial and parallel magnetized PMs,
the PM loss can be obviously reduced by only a few PM divisions; while for the Halbach magnetized PMs, the
segment number needs to be increased accordingly. HSPMM thermal performances are studied by the Lumped
method, while it is learnt increasing the cooling water volume flow rate is helpful to reduce machine operating
temperature in a certain range. HSPMM rotor strength analyses are also carried out. It is found the sleeve equivalent
stress is more obviously affected by the interference fit; increasing sleeve thickness is helpful to reduce the PM
tensile stress, especially when a large interference fit applied.

Declaration of competing interest


The authors declare that they have no known competing financial interests or personal relationships that could
have appeared to influence the work reported in this paper.

Acknowledgments
This work is supported by Ministry of Science & Technology, China under National Key R&D Program of China
(Grant No. 2021YFE0108600). And this work is also supported by National Natural Science Foundation of China
under grant 52077121, National Natural Science Foundation of China Key International Cooperation project under
grant 51920105011. This work is also supported by Liaoning Province Natural Science Foundation, China under
grant 2021-KF-24-01.

References
[1] Geest Martin, Polinder Henk, Ferreira Jan, Christmann Markus. Power density limits and design trends of high-speed permanent magnet
synchronous machines. IEEE Trans Transp Electrif 2015;1.3(2015):266–76.
[2] Hong Do-Kwan, Lee Tae-woo, Jeong Yeon-Ho. Design and experimental validation of a high-speed electric turbocharger motor
considering variation of the L/D ratio. IEEE Trans Magn 2018;54.11(2018):#2801904.
[3] Jang Gang-Hyeon, Ahn Ji-Hun, Kim Byung-Ok, Lee Dong-Hyun, Bang Jae-Sung, Choi Jang-Young. Design and characteristic analysis
of a high-speed permanent magnet synchronous motor considering the mechanical structure for high-speed and high-head centrifugal
pumps. IEEE Trans Magn 2018;54.11(2018):#8204906.
217
Y. Zhang, Z. Qi, H. Wang et al. Energy Reports 8 (2022) 209–218

[4] Yu Yanan, Liang Deliang, Liang Zhe, Ze Qiji. Calculation for stator loss of high-speed permanent magnet synchronous machine in
torque-speed envelope and restraint approach for circulating current in windings. CES Trans Electr Mach Syst 2018;2.2(2018):211–9.
[5] Huang Ziyuan, Fang Jiancheng, Liu Xiquan, Han Baocheng. Loss calculation and thermal analysis of rotors supported by active magnetic
bearings for high-speed permanent-magnet electrical machines. IEES Trans Ind Electron 2016;63.4(2016):2027–35.
[6] Dong Jianning, Huang Yunkai, Jin Long, Lin Heyun, Yang Hui. Thermal optimization of a high-speed permanent magnet motor. IEES
Trans Mag 2014;50.2(2014):#14128103.
[7] Fang Haiyang, Li Dawei, Qu Ronghai, Li Jian, Wang Cong, Song Bao. Rotor design and eddy current loss suppression for high-speed
machines with a solid-PM rotor. IEES Trans Ind Appl 2019;55.1(2019):448–57.
[8] Zheng Pei, Li Jian, Qu Ronghai, Fang Haiyang. Electromagnetic design issues of high-speed permanent magnet machine. In: XXII
international conference on electrical machines (ICEM). Lausanne, Switzerland; 2016, p. 1650–6.

218

You might also like