IEEE Electrification Magazine - June 2014 - 28

Thermal Circuit Model
6.5726

4.445 05
4.976 94
5.508 82
6.04071

3.381 2
8

Specific Electric Loading A (A/mm)

7
27
.8
10
58
.29
10
1
39
76
9.

2
32 0
9.2
14

2
22 24 0 6 00
0
00
18 20 0
0
00

0
8. 70

I E E E E l e c t r i f i c ati o n M agaz ine / j un e 2014

8
8.16

00
10

2
88
50
5.
94
76
4.9
05
45
4.4

3.913 17

80 0

110

7.10448 0
0
12

the thermal analysis used in the
motor design is usually based on two
16
14
types of models: the numerical field
00
00
100
model or the lumped-circuit model. a
90
lumped-circuit model may not pro9.
8. 70 232 02
vide much localized detail as a numer7.6
80
0
8.16
14
36
7.1
8 25
37
ical field model but will provide a good
0
6
4
.
572
70
48
6.0
6
40
global accuracy. a novel 3-D thermal
7
.6
3
6
3
7
71
7.1
60
6.572 6 04 48
lumped-circuit model has been devel5.508 8 6.040 71
50
oped to provide fast evaluation results
2
4
4
.9
5.508 82
00
3.3
4.445 0 76 94
4.976 94
81
3.913
for design changes that cause thermal
5
40
1
4.4
28
45 05
7
3.913 17
80 0
2.84
3.91317
performance variations.
3.
38
12
8
9
3.381 28
4
30
2.8494
2.317 51
60 0
2.8494
one of the most complex aspects
2.31
751
2.31751
1.78563
20
1.78563
1.78563
of
the
ipM thermal analysis is taking
1.253 74
1.25374
20 0 1.25374
care of the circumferential and axial
10
6
8
10
12
14
16
18
20
22
24
temperature differences. the convenArmature Current Density J (A/mm2)
tional lumped-circuit design method
cannot express the differences in conFigure 3. The trends of t peak with respect to q.
vection heat transfer and interface
thermal resistance. to solve this problem, the motor thermal model is divided into many circumferential and axial elements, and each element is modeled
Initial Values of
as a component in a lumped thermal circuit. as shown in
Parameters
Iteration
(P, kT, kh, kf, ks, kyĂ)
figure 5, the ipM has a water jacket with axial water channels in it. considering the different temperature distribution
along the circumferential directions, we have divided the
Peak Power Density
ipM circumferentially into four parts as shown in figure 6,
Optimization
and the lumped circuit for part 1 is shown in figure 7 as a
conventional lumped circuit. the black points represent the
power sources and the white points the power of convection
Range of Specific
Other Design
heat transfer. the winding is divided into inner and outer
Electric Loading and
Constraints
Current Density
layers to approximate the conductors in the stator slot. the
end windings and slot conductors are treated as separate
heat sources. in more complex applications, the heat transMagnetic Design
fer between the stator and rotor is considered.
Based on Magnetic Circuits
all of the ipM thermal network blocks are shown in
figure 8. each block is a complex thermal network as
shown in figure 7. the transient temperature rise of part 1
Thermal Design
is calculated using the coolant inlet temperature Tin as the
Based on Thermal Circuits
given condition. the solved coolant outlet temperature T2
is used to calculate the transient temperature rise of parts
2 and 4. the coolant outlet temperature T3 and T4 is used
Figure 4. The integrated circuit design process.
to calculate the transient temperature rise of part 3.
the following assumptions are made for the whole
thermal network:
xx
a large thermal resistance exists at the boundary of
the different parts, and this thermal resistance is used
to simulate the heat flow generated by the temperature difference.
xx
the different coolant inlet temperatures of each part
are used to simulate the temperature difference of the
circumferential water channel.
xx
considering the air flow inside the housing, the same
air-cooled components in the different parts have difFigure 5. A water jacket with axial water channels.
ferent convective heat transfer coefficients.
120

Table of Contents for the Digital Edition of IEEE Electrification Magazine - June 2014