Siemens PLM Software

Simcenter MAGNET
validation example
Simulating the continuous duty mode of a
linear servomotor

Executive summary
Modern permanent magnet linear synchronous motors (PMLSM) can be
applied to a wide range of projects because of their high efficiency, thrust
force density and acceleration. To ensure normal operation and maximum
thrust force density, the thermal performance must be done at the design
stage. Based on simulations using Simcenter MAGNET™ software, in this
paper we have coupled an electromagnetic-thermal finite element analy-
sis (FEA) of a 10-pole, 12-slot flat type iron core PMLSM. This motor was
the first linear servomotor designed and produced in Bulgaria and was
used as a main drive unit of a linear electromagnetic transportation sys-
tem for the light industry.

Dr. Peter Uzunov, Agile Magnetics Inc.

siemens.com/simcenter
 White paper | Simcenter MAGNET validation example

Abstract

In electrical machines, the design of heat transfer is as

important as the electromagnetic design of the machine
because the temperature rise forces it to be constantly
under load. The analysis of heat transfer in electric
motors is nonlinear and often more complex to predict
and simulate when coupling with the electromagnetic
behavior.
Linear motors are well known in the art and used in
many different industries. For example, they are used in
industrial transportation systems [1].
For the modern PMLSM, heat transfer is the principal
restrictive effect on the motor power.
The dynamics of linear motors depends on the movable
part mass and the generated electromagnetic force.
When the motor winding is part of the primary (moving
component), you can reduce total weight in the design
stage by considering higher values of current density.
This results in using a conductor of greater diameter,
hence the necessity for deeper and wider magnetic core
slots. In turn, this increases moving part weight and
worsens motor dynamics. On the other hand, the higher
values of current density result in increased Joule heat
losses in the winding, which in turn leads to accelerated
motor warming. If the motor doesn’t have a cooling
Figure 1: Prototype of permanent magnet linear synchronous
system, the temperature must not exceed the admis- motor PMLSM-3.
sible values as determined by the class of isolation
during the design process.
The linear motor being studied is shown in figure 1. It
All these considerations imply that reliable methods has a long magnet track composed of a series of alter-
must be available for calculating winding overheating. nating polarity Neodymium rare earth magnets
Modern thermal design techniques can be classified into mounted over a ferromagnetic back iron. This part is
two general methods; the analytical lumped circuit called the secondary. The moving part of the motor
approach and numerical analysis by finite element (called the primary) has a coil assembly, formed from
method (FEM). The FEM approach gives a more detailed multiple laminations. The coil assembly is wound
resolution throughout the machine and can be used to around teeth, which have slots between them.
model complex component shapes. The PMLSM has a 10-pole and 12-slot configuration.
For that reason, in this work FEM was used for analyz- The motor winding is double-layer concentrated. It was
ing the nonstationary temperature field of a PMLSM. supplied and controlled via a versatile servo drive
Thermal simulations of the investigated linear motor (VSD)-E 160 [2] and mounted over the primary’s core.
under normal work conditions were performed using The electrical energy is transferred to the moving part
this approach. via a linear planar transformer [3]. The VSD was

Siemens PLM Software 2

 White paper | Simcenter MAGNET validation example

controlled wirelessly from the personal computer (PC)

with the help of ZigBee industrial communication tech-
nology [3].
The motor does not have a cooling system for two
reasons:
• The magnetic core with wound coil is mounted on the
motor housing, which is an aluminum plate that is 10
millimeters (mm) thick. As the core is in tight contact
with the motor body and there isn’t any heat insula-
tion between them, the motor body performs the
function of a heat sink, dissipating the generated heat
• The winding is located on the movable part, which
moves 1.5 meters per second (m/s) and cools down
due to the motion

Siemens PLM Software 3

 White paper | Simcenter MAGNET validation example

FEM modeling

Since a coupled electromagnetic-thermal FEA problem The magnetic field distribution can be obtained after
was solved, the FEM modeling had two stages. First, we solving the equation (6) using the Simcenter MAGNET
built a model of linear motor for simulation in transient with motion solver [4] relative to the magnetic
Simcenter MAGNET [4]. field intensity H. The magnetic flux density B is com-
puted from the constitutive tensor relation (2). The
1. Electromagnetic modeling current density displayed in the postprocessor is just
Relevant equations curl of H.
In the conducting medium, one must start with the Because of the symmetry of this machine, it is possible
second Maxwell equation to find the general field to model only half of the motor, as one can see in figure
equation: 2. Modeling only half of the motor decreases the time
required for both meshing and simulation.
(1)
Alternating currents in the winding of the mover pro-
duce a traveling magnetic field. It interacts with the
where B is the magnetic field density, and E is the elec- magnetic field of the permanent magnets (PM) and
tric field intensity. linear motor work. Because the moving part has better
The following constitutive tensor relations are also cooling conditions for a linear motor, it is important we
used: simulate the real motion. For this reason, the motion
component Motion#1 part was built in the model.
(2)
The Motion#1 component travels inside one remesh
component named Remesh type A with assigned mate-
(3) rial air. A second remesh component named Magnetic
track remesh that was assigned material air was also
included in the model. Only the mesh in these remesh
Where J is the total current density given by: regions were changed during the transient with motion
(4) solution. A maximum element size of 2.5 mm was
assigned to these components to obtain a better mesh
and faster meshing time. This allows the user to mini-
In which it is assumed the permittivity tensor Ɛ does not mize the aspect ratios of the elements.
depend explicitly on time. The final model for electromagnetic simulation is shown
In equations (2)-(4) y is the conductivity and μ is the in figure 2.
permeability of the medium.
Using (3) with the first Maxwell’s equation:
(5)

and (2) in (1) yields to

(6)

Figure 2: 3D FEM model of PMLSM for electromagnetic simulations.

Siemens PLM Software 4

 White paper | Simcenter MAGNET validation example

Col 1 Col 2 Col 7 Col 8

T1 T2 T1 T2 T1 T1 T2 T2

AC

Figure 3: Selected virtual air components. Col 3 Col 4 Col 10 Col 11

T1 T2 T1 T2 T1 T2 T1 T2

The material properties of virtual air were assigned to AC

the mover air replacement and magnetic track virtual

air regions with the aim of improving the accuracy of
the Maxwell Stress force calculations. They are indi- 3
Col 5 Col 6 Col 11 Col 12
cated in red in figure 3. T1 T2 T1 T2 T1 T2 T1 T2

In this paper FEM was combined with circuit theory AC

because this approach works the best in analyzing these
electromagnetic devices. During the modeling of the
linear motor, all 12 coils are connected in a double-layer
concentrated winding scheme (figure 4) and inserted into
a circuit using the Simcenter MAGNET circuit tool [4]. Figure 5: Circuit diagram of motor winding.

This internal circuit is shown in figure 5.

S1 S2 S3 S4 S5 S6
The three phases of motor winding are supplied from
three current sources. These sources are sinusoidal and
simulate the digital drive VSD-E.
For best solution results a fine FE mesh was used, as
10

11

12
1

shown on figure 6. This 3D mesh includes 309,441

nodes and 1,715,087 tetrahedral FE.

A B C X Y Z
Figure 4: Scheme for motor winding.

Figure 6: Solution 3D mesh for time step t=70 ms.

Siemens PLM Software 5

 White paper | Simcenter MAGNET validation example

2. Thermal modeling The iron loss fields are obtained from the calculated B
To simplify thermal analysis of the PMLSM, basic field values and the Steinmetz equation. Iron losses are
assumptions are proposed as follows: provided as two separate fields: hysteresis and lamina-
tion eddy current. In Simcenter MAGNET 7.9 [4], it is
• Environmental temperature is maintained at 23
easy to model iron losses using the curve of losses
Celsius (°C) because thermal experiments were pro-
versus magnetic flux density based on the data provided
vided for this temperature
by material manufacturers (figure 7).
• Due to the relatively low temperature of the
machine’s outer surface, the radiation phenomenon is
neglected 3.4
3.0
Relevant equations
The nonstationary temperature field could be obtained
after solving the equation expressed for the three-

Loss (W/kg)
2.0
dimensional case in Cartesian coordinate system as
follows:
1.0
(7)

0
where:
0.1 0.5 1 1.5 1.7
qv – volumetric heat generation rate;
Bpeak (tesla)
kx , ky , kz – thermal conductivities in global X, Y, Z direc-
Figure 7: Iron losses versus flux density for used core laminations 5212,
tions, respectively; thickness 0.35 mm (GOST 21427.2).
T – temperature;
C - specific heat;
ρ - mass density;
t – time. Appropriate boundary conditions (BC) and initial condi-
tions are needed for the thermal FEA. In the present
The volumetric heat generation rate in winding was
work the convection BC was used. In Simcenter
calculated as follows:
MAGNET Thermal software [4], this is called a thermal
(8) environment (TE) boundary condition. This boundary
condition approximates the effect of an environmental
fluid in terms of heat convection and radiation.
where: PCu – copper losses, V – active linear motor
volume. Heat flux can be calculated as:

The copper losses have been calculated as: (11)

(9)
where:
where σ = 1/ρ is the conductivity. hc - heat transfer coefficient;
T - surface temperature;
To determine the resistance at a higher temperature, T∞ - ambient temperature.
the following well-known equation has been used:
To determine the 3D temperature field distribution in
(10) the linear motor, equation (7) is solved numerically with
the help of Simcenter MAGNET Thermal. For this pur-
pose, the transient 3D solver [4] was used.
where: R20- cooper wire resistance at initial temperature
T0=20ºC, a=0.0039 1/K – copper wire coefficient.

Siemens PLM Software 6

 White paper | Simcenter MAGNET validation example

Thermal FEM model

Each coil winding is composed of multiple turns of
American Wire Gage (AWG) wire, and its components
are modeled assuming a stranded structure formed
from the component’s assigned material (the primary
material is copper) and a secondary material (isolation).
Simcenter MAGNET Thermal approximates this micro-
structure as a homogeneous composite material having
anisotropic thermal conductivity and composite values
Figure 8: FEA model of PMLSM for a coupled electromagnetic-thermal solu-
of specific heat and mass density.
tion with applied boundary condition.
The materials of different motor parts as well as their
thermal performance parameters are given in table I.
Because the nonlinear problem was solved, the strong These environmental boundary conditions were applied
dependency of magnetic permeability and demagneti- to all the surfaces of the secondary.
zation for the permanent magnet material from the The primary of the linear motor travels with a linear
temperature must be considered [4]. speed of velocity (ν) = 1.5 m/s and the motor winding
was not potted in epoxy. Therefore, the outer surfaces
Boundary conditions of the ferromagnetic core, motor winding and alumi-
For thermal simulation of this linear motor a new num body are exposed to the cooling air and have
approach was used. The environmental boundary condi- better cooling conditions. Therefore, the speed 1.5 m/s
tions were assigned to the external surfaces of the two of forced convection was input in the environmental
virtual air components. This was done because of the boundary condition for mover air replacement.
complexity of the primary geometry and to save time
during model setup. The final model for coupled electromagnetic-thermal
simulation is shown in figure 8.
For all non-motion surfaces in the motor model, envi-
ronmental boundary conditions were used as follows:
• Heat transfer coefficient h0 = 10 [W/m2K]
• Ambient temperature T0 = 23ºC

Item Material k (W/m°C) ρ (kg/m3) C(J/kg°C)

Back iron Cold rolled 1010 steel 49.8 7600 448
Air gap Air k=ƒ(T) 1.205 C=ƒ(T)
Permanent magnets NdFeB N52 7.6 7500 460
Winding Cooper 386 8954 383.1
Mover core M-16 29 Ga 25 7650 490
Slot insulation Nomex 430 A 0.18 930 1340
Bodies of motor and magnet separators Aluminum 204 2707 896

Table 1: Material properties

Siemens PLM Software 7

 White paper | Simcenter MAGNET validation example

Motor part Power losses, watts

Winding 68.94
Permanent magnets and 6.8
magnet separators
Iron core 3.46
Figure 9: Distribution of magnetic flux density field in primary’s magnetic
core, secondary’s back iron and PMs for time step t=70 ms. Table 2: Time average power losses

Results from electromagnetic and thermal FEA Figure 11 shows the temperature of main motor parts
The coupled problem was solved on a 128-gigabyte versus time. In figure 12 the distribution of total heat
(GB) random access memory (RAM) workstation based source density is shown for the last time step.
on two Intel Xeon central processing units (CPUs) work-
ing in parallel.
70
The time average values of power losses in the main
motor parts are given in table 2. 60

In figure 10 the distribution of motor temperature field 50

was shown for three time steps: Temperature, C
40

30

20

10

0
0 500 1000 1500 2000
Time, s

(a) Winding Core Permanent magnet Separator Slot isolation

Figure 11: Maximum temperature versus time for motor parts.

(b)

Figure 12: Distribution of total heat sources density for time step t=1920s.

(c)

Figure 10: Nonstationary temperature field distribution for time steps:

a) t= 192 s; b) t= 960 s; c) t= 1920 s.

Siemens PLM Software 8

 White paper | Simcenter MAGNET validation example

Experimental validation of results from

thermal FEA

No. Quantity Value

1 Number of phases 3
2 Number of poles 10
3 Slots number 12
4 Rated phase voltage (V) 70 V
5 Rated phase current (A) 8 ampere (A)
6 Efficiency 0.8
7 Power factor 0.9
8 Air gap 1 mm
9 Pole pitch 15 mm
10 Tooth pitch 12.5 mm
11 Thrust 185 Newton (N)
12 Speed 1.5 m/s
13 Mover mass 6 kilograms (kg)

Figure 13: Experimental test setup for linear servomotor PMLSM-3. Table 3: Specifications of the test bench

To verify the results of FEA for the thermal field, one The main technical specifications of the test linear servo-
thermal experiment was conducted. motor are shown in table III.
The laboratory set system was equipped with a prototype Once switched on, the motor heats up. Measuring and
of the investigated linear servomotor PMLSM-3. The recording the temperature was done with LabVIEW[7].
experimental setup is shown in figure 13.

Siemens PLM Software 9

 White paper | Simcenter MAGNET validation example

(a)

Figure 15: Temperature versus time plot from LabVIEW.

70

60

50
Temperature, C

40

30

20 FEA
10
Experiment
(b) 0
0 500 1000 1500 2000 2500
Figure 14: Program [RT] Cont Acq (SAR).vi.
Time, s
Figure 16: Comparing experimentally and FEA obtained temperature values
for motor winding.
In its environment, the special program [RT] Cont Acq
(SAR).vi was developed. Its contents are shown in figure
14 a and b. As the plot shows, there is a good correlation between
the results from the thermal FEA and the experimental
With the help of this program, the experimental relation results.
temperature/time Т=ƒ(t) for the winding is obtained
directly (see figure 15). From both the simulations and the experimental thermal
study, we find the heating of the permanent magnets is
In the figure 16, the temperature versus time results weak and does not change their magnetic properties
obtained by both experiment and Simcenter MAGNET FEA because their maximum temperature is 80 °C.
are compared.
Motor winding heating is within the normal range defined
by the slot isolation and the AWG wire class.

Siemens PLM Software 10

 White paper | Simcenter MAGNET validation example

Conclusion

Simcenter MAGNET was used to perform electromag-

netic and thermal simulations of a PMLSM in continuous
duty mode. This approach facilitates calculating tem-
perature distribution and selecting proper current den-
sity for motor winding.

References
1. Gieras, J.F., Zbigniew J.Piech, Linear Synchronous Motors:
Transportation and Automation Systems, second edition, Boca Raton,
Florida, USA: CRC Press LLC, 2000.
2. Kotkanen, T., “Granite Device Knowledge Wiki,” available online: http://
granitedevices.com/wiki/VSD-E_and_VSD-XE
3. Uzunov, P.M., “Linear electromagnetic transportation system,”
Mechatronica SC, Technical Report on Contract NIF-02-22 / 15.12.2012 /
Executive Agency for Promotion of Small and Medium-sized Enterprises,
Gabrovo, Bulgaria, 2015.
4. https://support.mentor.com
5. Shastlivii, G., “Heating of induction motors,” Kiev Science community,
1966.
6. Uzunov, P.M., Banovich, Mladen, “Design Simulations and Force Analysis
of Permanent Magnet Linear Synchronous Motor,” International Review
on Modeling and Simulations (IREMOS), Vol.10, No. 1, February 2017.
7. www.ni.com

Siemens PLM Software 11

Siemens PLM Software About Siemens PLM Software
Siemens PLM Software, a business unit of the Siemens
Headquarters Digital Factory Division, is a leading global provider of
Granite Park One software solutions to drive the digital transformation of
5800 Granite Parkway industry, creating new opportunities for manufacturers
Suite 600 to realize innovation. With headquarters in Plano, Texas,
Plano, TX 75024 and over 140,000 customers worldwide, Siemens PLM
USA Software works with companies of all sizes to transform
+1 972 987 3000 the way ideas come to life, the way products are real-
ized, and the way products and assets in operation are
Americas used and understood. For more information on Siemens
Granite Park One PLM Software products and services, visit
5800 Granite Parkway siemens.com/plm.
Suite 600
Plano, TX 75024
USA
+1 314 264 8499
Acknowledgements
Europe The author expresses his thanks to James Atwood,
Stephenson House president of Agile Magnetics Inc., Colin Larose, PhD,
Sir William Siemens Square from the Siemens customer support team and Gilles
Frimley, Camberley Fillion, PhD, from the Siemens presales consulting team
Surrey, GU16 8QD for their help on this white paper.
+44 (0) 1276 413200

Asia-Pacific
Unit 901-902, 9/F
Tower B
Manulife Financial Centre
223-231 Wai Yip Street
Kwun Tong, Kowloon
Hong Kong
+852 2230 3333

siemens.com/plm
©2019 Siemens Product Lifecycle Management Software Inc. Siemens and the Siemens logo
are registered trademarks of Siemens AG. Femap, HEEDS, Simcenter, Simcenter 3D,
Simcenter Amesim, Simcenter FLOEFD, Simcenter Flomaster, Simcenter Flotherm, Simcenter
MAGNET, Simcenter Motorsolve, Simcenter Samcef, Simcenter SCADAS, Simcenter STAR-
CCM+, Simcenter Soundbrush, Simcenter Sound Camera, Simcenter Testlab, Simcenter
Testxpress and STAR-CD are trademarks or registered trademarks of Siemens Product Lifecycle
Management Software Inc. or its subsidiaries in the United States and in other countries. All
other trademarks, registered trademarks or service marks belong to their respective holders.
76930-A5 4/19 Y

12