Electric Motor Thermal

Management Research
Annual Progress Report
Kevin Bennion

NREL is a national laboratory of the U.S. Department of Energy

Office of Energy Efficiency & Renewable Energy
Operated by the Alliance for Sustainable Energy, LLC
This report is available at no cost from the National Renewable Energy
Laboratory (NREL) at www.nrel.gov/publications.

Management Report
NREL/MP-5400-67121
October 2017

Contract No. DE-AC36-08GO28308

 Electric Motor Thermal
Management Research
Annual Progress Report
Kevin Bennion

NREL is a national laboratory of the U.S. Department of Energy

Office of Energy Efficiency & Renewable Energy
Operated by the Alliance for Sustainable Energy, LLC
This report is available at no cost from the National Renewable Energy
Laboratory (NREL) at www.nrel.gov/publications.

National Renewable Energy Laboratory Management Report

15013 Denver West Parkway NREL/MP-5400-67121
Golden, CO 80401 October 2017
303-275-3000 • www.nrel.gov
Contract No. DE-AC36-08GO28308
 NOTICE

This report was prepared as an account of work sponsored by an agency of the United States government.
Neither the United States government nor any agency thereof, nor any of their employees, makes any warranty,
express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of
any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately
owned rights. Reference herein to any specific commercial product, process, or service by trade name,
trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation,
or favoring by the United States government or any agency thereof. The views and opinions of authors
expressed herein do not necessarily state or reflect those of the United States government or any agency thereof.

Cover Photos by Dennis Schroeder: (left to right) NREL 26173, NREL 18302, NREL 19758, NREL 29642, NREL 19795.

NREL prints on paper that contains recycled content.

Electric Motor Thermal Management Research
Principal Investigator: Kevin Bennion
National Renewable Energy Laboratory
Transportation and Hydrogen Systems Center
15013 Denver West Parkway
Golden, CO 80401
Phone: (303) 275-4447
E-mail: kevin.bennion@nrel.gov

DOE Technology Development Manager: Susan A. Rogers

U.S. Department of Energy
1000 Independence Ave. SW EE-3V
Washington, DC 20585
Phone: 202-586-8997
E-mail: susan.rogers@ee.doe.gov

NREL Task Leader: Sreekant Narumanchi

Phone: 303-275-4062
Email: sreekant.narumanchi@nrel.gov

Contractor:
Contract No.:

Abstract/Executive Summary
With the push to reduce component volumes, lower costs, and reduce weight without sacrificing performance
or reliability, the challenges associated with thermal management increase for power electronics and electric
motors. Thermal management for electric motors will become more important as the automotive industry
continues the transition to more electrically dominant vehicle propulsion systems. The transition to such
systems leads to higher-power duty cycles for electric-drive systems. Thermal constraints place significant
limitations on how electric motors ultimately perform, and as thermal management improves, there will be
direct tradeoffs among motor performance, efficiency, cost, and the sizing of electric motors to operate within
the thermal constraints.

The goal of this research project is to support broad industry demand for data, analysis methods, and
experimental techniques to improve and better understand motor thermal management. Work in FY16 focused
on two areas related to motor thermal management: passive thermal performance and active convective
cooling. Passive thermal performance emphasized the thermal impact of materials and thermal interfaces
among materials within an assembled motor. Active convective cooling focused on measuring convective heat-
transfer coefficients using automatic transmission fluid (ATF).

Accomplishments
● Completed construction of experimental test apparatus to measure large-scale variation in ATF
convective heat transfer coefficients
● Completed measurement of fan jet nozzle for ATF heat transfer in collaboration with industry partner
● In collaboration with Oak Ridge National Laboratory (ORNL), performed measurements and analysis
of slot winding materials and submitted a manuscript summarizing the results to a journal for potential
publication
● In collaboration with Ames Laboratory, performed measurements of mechanical and thermal
properties of new magnet materials.

    
1
Introduction
Thermal management for electric motors is important as the automotive industry continues to transition to
more electrically dominant vehicle propulsion systems. With the push to reduce component size, lower costs,
and reduce weight without sacrificing performance or reliability, the challenges associated with thermal
management for power electronics and electric motors increase. The transition to more electrically dominant
propulsion systems leads to higher-power duty cycles for electric drive systems. Thermal constraints place
significant limitations on how electric motors ultimately perform. As summarized by Thomas Lipo, “[a]n
optimized thermal design can help increase machine rated power substantially, almost without any increase of
its manufacturing costs.” [1]. The performance limitations caused by motor heating are highlighted in Figure
I-1. The motor's ability to increase running time at higher power levels within electrical operating limits is
directly related to the ability to remove heat from critical components. As thermal management improves, there
will be a direct tradeoff among motor performance, efficiency, cost, and the sizing of electric motors to operate
within the thermal constraints.

Figure I-1: Thermal management impact on motor performance to support increased power
Image Source: NREL

Thermal management of electric motors is a complex challenge because of the multiple heat transfer paths
within the motor and the multiple materials and thermal interfaces through which the heat must pass to be
removed. The technical challenges to motor thermal management are summarized by Hendershot and Miller as
follows: “Heat transfer is as important as electromagnetic and mechanical design. The analysis of heat transfer
and fluid flow in motors is actually more complex, more nonlinear, and more difficult than the electromagnetic
behavior” [2]. Figure I-2 provides a cut cross-section view illustrating heat transfer and cooling paths for
automotive traction drive applications. The heat generated by the electric motor is distributed throughout
multiple components within the electric motor. For example, heat is generated due to losses within the stator
slot-windings, stator end-windings, stator laminations, rotor laminations, and rotor magnets or conductors. The
distribution of the generated heat within the components is dependent on the motor type and the operating
condition (torque/speed) of the motor. The selected cooling approach for the motor impacts the path of heat
flow through the motor and the temperature distribution of components. For example, as shown in Figure I-2, a
motor cooled with a stator cooling jacket will require heat generated within the slot windings to pass through
multiple material layers and material interfaces before the heat is extracted through the cooling jacket. The
thermal properties of the materials and the thermal contact resistances due to the material interfaces impact the
temperature distribution inside the motor as heat flows into the cooling jacket. Alternatively, direct cooling of
the windings with oil or ATF reduces the heat transfer path from the motor windings to the coolant. However,
heat from the stator must pass through several interfaces. The resulting changes in the temperature distribution
within the motor lead to hot spots within the motor that could be difficult to measure.

2
Figure I-2: Heat must pass through several layers within the motor to be extracted through active cooling such as through a
cooling jacket or spray cooling with oil such as ATF.
Image Source: NREL

Thermal management of the motor is not only important for the reliability of the motor, but the temperatures of
the components within the motor affect material properties that directly relate to the torque production, control,
and efficiency of the motor. For this reason, motor designers need accurate thermal models of the electric
motor during the design and control development of the motor. Critical to the ability to accurately model the
thermal behavior of the motor is access to data describing critical thermal characteristics of the motor. Such
data include direction-dependent thermal conductivity measurements of nonuniform motor components such as
lamination stacks and windings. It also includes data to quantify thermal contact resistances between
components in the motor. Finally, it includes data to support the modeling and design of active cooling of the
motor and the convective heat transfer coefficients possible from alternative cooling approaches.

Approach
The ability to remove heat from an electric motor depends on the passive stack thermal resistance within the
motor and the convective cooling performance of the selected cooling technology. In addition, as new
materials are developed, it is important to characterize temperature-dependent material mechanical properties
and thermal properties. Characterization of new materials enables motor designers to evaluate the potential
performance tradeoffs of new materials for motor applications. For this reason, the approach for the research
project splits the efforts between three primary categories as illustrated in Figure I-3.

Figure I-3: Approach to motor thermal management divided between passive thermal design and active convective cooling
Photo Credits: Doug DeVoto, Jana Jeffers, Kevin Bennion, NREL

3
Passive thermal design refers to the geometrical layout, material selection, and thermal interfaces that affect
the heat-spreading capabilities within the motor. The ability for heat to spread through the motor affects the
thermal temperature gradients within the motor. The active convective cooling technology is the cooling
mechanism that ultimately removes the heat from the motor and transfers the heat to another location to reject
the heat to the ambient environment. The material characterization work focuses on the measurement of
mechanical and thermal properties, in collaboration with project partners, of new individual materials relevant
to motor applications.

Active cooling
The two common approaches highlighted in Figure I-2 for active cooling include: 1) directly cooling the motor
with ATF, and 2) cooling the motor with a cooling jacket surrounding the stator. The advantages of either
cooling approach depend on the application's coolant availability, the motor geometry, the motor winding
configuration, and the motor loss distribution. The advantage of cooling using ATF is that it is possible to
directly cool the motor windings or rotor. Past work focused on measurement of average convection
coefficients of ATF jets passing through a circular orifice directly impinging on target surfaces representative
of motor end windings. In the area of active cooling, the focus during FY16 emphasized an alternative spray
pattern and spatial mapping of the heat transfer coefficients at the stator winding scale.

The heat transfer coefficient will vary along the larger-scale stator end-winding illustrated in Figure I-4. As the
ATF flows over the end-winding, the heat transfer will be different than around the impingement zone of the
jet. The irregular surface caused by the wire bundles also complicates the fluid flow paths and the heat transfer.
During FY16, the National Renewable Energy Laboratory (NREL) finished construction of an experimental
setup and enclosure to measure the heat transfer variation on the motor end-winding. The initial experiments
exclude the irregular surfaces caused by the wire bundles seen in Figure I-4, but additional end-winding
geometry complexity can be incorporated in future experiments using the same experimental setup. The
following sections summarize the design and construction of the experimental equipment and the experiments
are part of ongoing work.

Figure I-4: Heat transfer due to ATF jet impingement will vary over the end-winding surface.
Photo Credit: Kevin Bennion, NREL

Passive Thermal Design

The passive thermal stack elements illustrated in Figure I-2 and Figure I-5 are critical to designing effective
thermal management systems for electric motors. The work supports improved thermal models for motor
design, but it also enables analysis to compare the potential impacts of new materials, fabrication methods, or
material processing on motor heat transfer. Figure I-5 illustrates a few of the critical elements that influence
the passive cooling of the motor or the ability of heat to flow through the motor. A few of the items highlighted
in Figure I-5 include the stator-to-case thermal contact resistance, lamination through-stack and in-plane
thermal conductivity, winding cross-slot thermal properties and thermal interface resistance between ground
insulation materials and the respective motor elements in contact with the slot liner or ground insulation.

4
Efforts continued in FY16 to measure passive stack elements within the motor as highlighted above in
collaboration with industry, universities, and ORNL.

Figure I-5: Passive stack thermal properties and terminology.

Image Source: NREL

Material Characterization
The material characterization work focuses on the measurement of mechanical and thermal properties of new
individual materials relevant to motor applications in collaboration with project partners. For FY16, the work
focused on collaborations with Ames Laboratory and its work to develop new magnet materials relevant for
electric motor applications. NREL provided support to Ames for mechanically and thermally characterizing
the new magnet materials. Tests focused on comparing the transverse rupture strength and thermal
conductivity of the new magnet materials with currently available materials. The test fixture for the transverse
rupture testing is shown in Figure I-6. The transverse rupture strength (TRS) is determined by Equation 1 [3]
using the force required to rupture the specimen (P), the distance between supports (L), the specimen width
(w), and the specimen thickness (t).

Figure I-6: Transverse rupture test apparatus.

Photo Credit: Doug DeVoto, NREL

3𝑃𝑃𝑃𝑃 (1)
𝑇𝑇𝑇𝑇𝑇𝑇 =
2𝑡𝑡 2 𝑤𝑤

5
Results and Discussion

The discussion included below is separated into the three main focus areas described above. The first section
summarizes progress for the material characterization results. The second section focuses on passive cooling of
the electric motor. The third section summarizes progress on active cooling with emphasis on using ATF for
cooling electric motors.

Material Characterization
Measurements of both transverse rupture strength and thermal conductivity were made, in collaboration with
researchers at Ames Laboratory, on new AlNiCo magnets developed at Ames. The transverse rupture strength
tests were performed at NREL over a range of temperatures from -40°C to 150°C. The test fixture inside the
environmental chamber is shown in Figure I-7. The tests were performed both on the new Ames-developed
material and two commercially available AlNiCo-based magnets (AlNiCo 8HE and AlNiCo 9). The results of
the transverse rupture strength are shown in the left image of Figure I-8. The new Ames material demonstrated
higher transverse rupture strength at each of the tested temperatures. The significance is a mechanically
stronger magnet that is easier for manufacturers to implement into motor designs.

Figure I-7: Magnet materials (left) being tested for transverse rupture strength within a thermal chamber (right).
Photo Credit: Doug DeVoto, NREL

The thermal conductivity of the new Ames-developed magnet was also measured in comparison with two
commercially available materials. The thermal conductivity measurements were performed at room
temperature by measuring the thermal diffusivity by the xenon flash technique. For the tested samples, the
average thermal conductivity of the Ames material was lower than the two commercial materials that were
evaluated. The significance of the lower thermal conductivity in a motor application was not evaluated at this
time.

6
Figure I-8: Transverse rupture test results of Ames material and commercial materials (left). Thermal conductivity test results
of Ames material compared with commercially available materials (right).
Image Source: NREL

Passive Thermal Design

The work on the passive stack thermal design focused on evaluating methods for measuring thermal properties
of composite wire bundle samples held together with varnish. Figure I-9a highlights examples of wire bundles
for the slot winding and end winding of the motor stator. The ability to measure the apparent thermal
conductivity of the wire bundle samples will enable the ability to evaluate the impact of new motor winding
materials. Having data on the performance impact of new materials will enable motor designers to evaluate the
impact in particular motor designs.

During FY16, NREL worked in collaboration with ORNL to focus more closely on the winding materials
within the electric motor. Figure I-9b and Figure I-9c show examples of the motor winding composite material
samples prepared by ORNL. The samples were prepared to measure the apparent thermal conductivity parallel
and perpendicular to the axis of the wire. ORNL and NREL compared three methods for measuring the
thermal conductivity using laser flash, transient plane source, and transmittance test methods [4]–[6]. Each of
the test methods was first compared by estimating the thermal conductivity of monolithic, isotropic, and
homogenous soda lime silicate glass (low thermal conductivity) and polycrystalline silicon carbide (high
thermal conductivity). The apparatus using the transmittance test method [6] with the composite wire bundle
sample is shown in Figure I-9d.

Figure I-9: Sample wire bundles in motor slot and end winding (a), winding sample blocks prepared by ORNL for thermal
property measurements perpendicular to wire axis (b), winding sample blocks prepared by ORNL for thermal property
measurements parallel to wire axis (c), test sample under test using transmittance test method (d).
Photo credits: Kevin Bennion, Emily Cousineau, NREL

7
In addition to performing the sample measurements, an effort was made to compare the experimental results
with derived estimates based on finite element analysis (FEA) models of the wire bundle. For the FEA model,
the copper fill factor of the sample was determined using two methods. One method determined the fill factor
for each individual sample using image analysis of the visible cross section of the wires. The second method
determined the fill factor by measuring the density of the test sample. Figure I-10 compares the experimental
results and the FEA estimates for two wire-size samples. The FEA Low and FEA High represent the range for
the low and high estimates for the copper fill factor of the sample. The measurement uncertainty of the data
was determined with the 95% confidence interval including all known random and systematic sources of error
[7].

Figure I-10: Transmittance test method results for thermal conductivity perpendicular to wire axis, and comparison of
experimental data with FEA model results based on low and high estimates for the sample copper fill factor for 22 AWG (left)
and 19 AWG (right).
Image Source: NREL

As seen in Figure I-10, the measured thermal conductivity values are larger than the predicted values. The FEA
modeling assumes ideal varnish fill (no voids) and no thermal interfacial losses between the materials within
the wire bundle. With these assumptions the FEA estimates would be expected to be higher than the measured
data. However, the FEA model also assumes ideal or uniform packing of the wires with the sample which is
not representative of the tested samples. The non-uniform packing of the wires could produce preferential
pathways for heat transfer resulting in higher measured thermal conductivity. Full details of the testing and
discussion of the results were submitted during FY16 for consideration of publication to the ASME Journal of
Thermal Science and Engineering Applications.

Active Cooling
Work on active convective cooling during FY16 focused on using ATF for cooling electric motors. Past work
measured the average convective heat transfer coefficients of circular orifice ATF jets directly impinging on
stationary target surfaces with surface features representative of motor end-windings [8]. A schematic of the
orifice jet test is shown in Figure I-11a. An alternative ATF jet geometry was tested during FY16, as shown in
in Figure I-11b. Figure I-11b shows a flat planar fan jet that impacts the target surface for measuring heat
transfer. The heat transfer coefficient is determined using the same approach as described in prior published
papers [8]. A key difference in the test is that unlike the orifice jet, the ATF for the planar fan jet is not limited
to impingement on the target surface.

8
Figure I-11: Test setup for circular orifice ATF jet (a). Test setup for planar fan jet (b).
Image Source: NREL

The assembled test apparatus is shown in Figure I-12. The jet nozzle-to-target distance was fixed at 25.4 mm.
The spray angle of the fan jet was dependent on the flow rate through the nozzle, as shown in Figure I-12.
Higher flow rates result in a larger spray angle. Depending on the spray angle, a portion of the ATF does not
impinge on the 12.7-mm-diameter target surface used to measure heat transfer. For this reason, only a portion
of the fluid is used to remove heat. The potential advantage of the planar fan jet is the ability to remove heat
over a larger surface, such as a motor end winding.

Figure I-12: Flow rate of 0.6 L/min led to 58° spray angle (left). Flow rate of 1.0 L/min led to 78° spray angle (right).
Photo Credit: Xuhui Feng, NREL

The results of the planar fan jet test results are shown in the left graph of Figure I-13. The heat transfer data
can be compared with the data published for the orifice nozzle [8], shown in the right graph of Figure I-13. For
the same volumetric flow rate through the nozzle, the heat transfer coefficient averaged over the 12.7-mm-
diameter target is higher for the orifice nozzle as compared to the planar fan jet. However, as mentioned
previously, not all of the ATF impinges on the target surface for the planar fan jet. The advantage of the planar
fan jet is the ability to impinge over a larger area. Figure I-14 shows graphs comparing the heat transfer
coefficient versus the fluid parasitic power, which includes the volumetric flow and the pressure drop. As seen
in Figure I-14, the difference in the heat transfer coefficient for the two tested nozzles is not as large for
equivalent parasitic power.

9
Figure I-13: Average heat transfer coefficient versus flow rate measured over 12.7-mm-diameter target surface for planar fan
jet (left) and orifice jet (right).
Image Source: NREL

Figure I-14: Average heat transfer coefficient versus parasitic power measured over 12.7-mm-diameter target surface for
planar fan jet (left) and orifice jet (right).
Image Source: NREL

The ability to measure the spatial variation of the heat transfer coefficient over a larger area required
construction of a larger test section for stator-scale ATF thermal measurements on motor end-windings. The
image on the left of Figure I-15 illustrates the goals of the experimental setup. The experiment was designed to
enable the measurement of heat transfer on multiple surfaces of the motor end winding while allowing for the
relative position between the heat transfer sensor and fluid jet impingement zone to change. Unlike prior
measurements, the fluid jet is not constrained to be in a fixed location relative to the heat transfer sensor
location. The experimental setup allows for the study of nozzle location, nozzle type, jet interactions, flow
rates, gravity, and alternative cooling designs along the inner diameter, outside diameter, and outside edge of
the motor end winding. The image on the right of Figure I-15 shows the operating experimental setup with the
orifice type nozzle spraying ATF on the heat transfer sensor installed in the motor end winding. The
comparison of the orifice type jet and the planar jet is part of ongoing work.

10
Figure I-15: Illustration of sensor package installed in end-winding (left), sensor package installed in motor and ATF impinging
on target surface (right).
Photo Credit: Kevin Bennion, NREL

Conclusions and Future Directions

Past work in the area of active convective cooling provided data on the average convective heat transfer
coefficients of circular orifice ATF jets impinging on stationary targets intended to represent the wire bundle
surface of the motor end-winding. The work during FY16 focused on the impact of alternative jet geometries
that could lead to improved cooling over a larger surface of the motor winding. Results show that the planar jet
heat transfer coefficients over a small (12.7-mm-diameter) target surface are not too much lower than the
circular orifice jet in which all of the ATF from the jet impinges on the target surface. The planar jet has the
potential to achieve higher heat transfer over a larger area of the motor end winding. A new test apparatus was
constructed to measure the spatial dependence of the heat transfer relative to the jet nozzle over a larger area
representative of a motor end-winding. The tested planar flow geometry has the potential to provide more
uniform cooling over the full end-winding surface versus the conventional jet configuration. The data will be
used by motor designers to develop thermal management strategies to improve motor power density.

The area of passive thermal design saw the completion of work in collaboration with ORNL to measure the
thermal conductivity of wire bundle samples representative of end-winding and slot-winding materials.
Multiple measurement techniques were compared to determine which was most suitable for measuring
composite wire bundle samples. NREL used a steady-state thermal resistance technique to measure the
direction-dependent thermal conductivity. The work supported new interactions with industry to test new
materials and reduce passive-stack thermal resistance in motors, leading to motors with increased power
density.

The work in the area of material characterization was performed in collaboration with Ames Laboratory. The
work focused on measuring the transverse rupture strength of new magnet materials developed at Ames. The
impact of the improved transverse rupture strength is a mechanically stronger magnet that is easier for
manufacturers to implement into motor designs. The thermal conductivity of the new magnet materials was
also measured in comparison to two commercially available AlNiCo magnet materials. The impact of the
thermal conductivity of the magnet material would need to be analyzed in the context of a motor application.

FY 2016 Presentations/Publications/Patents
1. Bennion, K. “Power Electronics Thermal Management Research.” 2016 DOE VTO Annual Merit
Review, Washington D.C., June 2016.

11
Acknowledgments
The author would like to acknowledge the support provided by Susan Rogers, Technology Development
Manager for the Electric Drive Technologies Program, Vehicle Technologies Office, U.S. Department of
Energy Office of Energy Efficiency and Renewable Energy.

The significant contributions from Emily Cousineau, Xuhui Feng, Gilbert Moreno, and Doug DeVoto (NREL)
to the project are acknowledged. The support, collaboration, and sample motor materials provided by Andrew
Wereszczak and Tim Burress (ORNL) are acknowledged as well. The support and collaboration with Emma
White, Liangfa Hu, and Iver Anderson of Ames Laboratory are also acknowledged.

References

[1] T. A. Lipo, Introduction to AC Machine Design, 3rd ed. Wisconsin Power Electronics Research
Center, University of Wisconsin, 2007.
[2] J. R. Hendershot and T. J. E. Miller, Design of Brushless Permanent-Magnet Motors. Oxford, UK:
Magna Physics Publishing, 1994.
[3] “Standard Test Method for Transverse Rupture Strength of Powder Metallurgy (PM) Specimens.”
ASTM B524-12, ASTM International, West Conshohocken, PA.
[4] “Standard Test Method for Thermal Diffusivity by the Flash Method.” ASTM E1461, ASTM
International, West Conshohocken, PA, 2013.
[5] “Transient Plane Heat Source (Hot Disc) Method.” ISO 22007-2, International Organization for
Standardization, Geneva, Switzerland, 2008.
[6] “Standard Test Method for Thermal Transmission Properties of Thin Thermally Conductive Solid
Electrical Insulation Materials.” ASTM D5470, ASTM International, West Conshohocken, PA, 2012.
[7] R. H. Dieck, Measurement Uncertainty: Methods and Applications. ISA, 2007.
[8] K. Bennion and G. Moreno, “Convective Heat Transfer Coefficients of Automatic Transmission Fluid
Jets with Implications for Electric Machine Thermal Management,” in ASME 2015 International Technical
Conference and Exhibition on Packaging and Integration of Electronic and Photonic Microsystems and ASME
2015 12th International Conference on Nanochannels, Microchannels, and Minichannels, San Francisco, CA,
United States, 2015.

12