Final Technical Report

Project Title: Ultra-Efficient and Power Dense Electric Motors for U. S. Industry

DOE Award Number: DE-FG36-08GO18132

Project Period: 9/30/2008 – 12/31/2012

Technical Contact: Rich Schiferl

26391 Curtiss Wright Parkway, Suite 102
Richmond Hts, OH 44143
(216) 261-3644 x211
(216) 261-3887 (fax)
Rich.Schiferl@baldor.abb.com

Recipient Organization: Baldor Electric

6040 Ponders Court
Greenville, SC 29615

Partners: Stephen D. Umans

Customers & End Users:

(free consultations provided)
Colfax Americas
Duke Energy Carolinas, LLC
DuPont Engineering
Howden / Buffalo Fan
Ameren
PeopleFlo

Date of Report: March 12, 2013

1
Acknowledgment, Disclaimer and Proprietary Data Notice –
DOCUMENT AVAILABILITY

Acknowledgment: “This report is based upon work supported by the U. S.

Department of Energy under Award No. DE-FG36-08GO18132“.

Disclaimer:

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. Any
findings, opinions, and conclusions or recommendations expressed in this report are those of the
authors and do not necessarily reflect those of the United States Government or any agency
thereof.

Document Availability: Reports are available free via the U.S. Department of Energy (DOE)
Information Bridge Website: http://www.osti.gov/bridge

Reports are available to DOE employees, DOE contractors, Energy Technology Data
Exchange (ETDE) representatives, and Informational Nuclear Information System (INIS)
representatives from the following source:

Office of Scientific and Technical Information

P.O. Box 62
Oak Ridge, TN 37831
Tel: (865) 576-8401
FAX: (865) 576-5728
E-mail: reports@osti.gov
Website: http://www.osti.gov/contract.html

2
Table of Contents

Section Title Page

1.0 Executive Summary 6
2.0 Introduction 7
3.0 Background 8
4.0 Results and Discussion 10
4.1 Technical approach and hypothesis guiding this approach 10
4.2 Experimental methodology, test procedures, characterization methods 15
4.3 Presentation and discussion of results 18
5.0 Benefits Assessment 27
6.0 Commercialization 30
7.0 Accomplishments 31
8.0 Conclusions 33
9.0 Recommendations 33
10.0 References 33

List of Acronyms /Terms

DOL – Direct on Line (full voltage, full frequency, three phase line operation)

LSIPM – Line-Start, Interior Permanent-Magnet (a type of motor which uses a cage winding
to allow DOL starting with a pull-in to synchronism at the end of the starting process)

NEMA – National Electrical Manufacturers Association – publisher of motor and generator

standards (such as NEMA MG1)

NEMA Design B – A general purpose motor type described (in NEMA MG1) by
characteristics including starting torque and starting current

NEMA Premium® Efficient – A level of efficiency described by NEMA MG1 which is the
highest efficiency level in this industry standard

NEMA Efficiency Bands – A geometric progression of efficiency levels defined in NEMA

MG1. Each higher band represents approximately a 10% reduction in motor losses. See
Table IV.

3
List of Figures
Figure Title Page
1 Typical NEMA Premium Efficiency Induction Motor 9
2a Rotor configuration with radial magnetization and uniform cage slot pattern 10
2b Rotor configuration with azimuthal magnetization and cage pattern 10
interrupted by magnets
3 Rotor magnet and cage pattern for many of the motors tested in this project 11
4 Open circuit flux pattern for the LSIPM motor of Fig. 3 11
5 Rotor showing end ring of die cast cage 12
6 Rotor die casting manufacturing cell 12
7 Rotor core stacked with magnets ready for die casting 13
8 Shaft insertion following die casting of rotor 13
9 Wound stator for use in prototype testing 14
10 Typical high-temperature rare-earth magnet second quadrant B-H curves 14
11 Magnetization of a prototype rotor 15
12a Dynamometer test setup for 30 – 50 HP steady-state tests 17
12b Data acquisition for steady-state testing 17
13 Test rig for starting and synchronization evaluation 18
14 30 hp tests: Efficiency vs voltage 19
15 30 hp tests: Current vs voltage 19
16 30 hp tests: Power factor vs voltage 20
17 380 V tests: Efficiency vs power 20
18 380 V tests: Current vs power 20
19 380 V tests: Power factor vs power 21
20 Speed during DOL starting at 380 V, 60 Hz, 113 Nm load torque, with a 24
coupled load inertia of 100 lb-ft2
21 Current during DOL starting at 380 V, 60 Hz, 113 Nm load torque, with a 24
coupled load inertia of 100 lb-ft2 - (a) Time waveform, (b) rms value of the
current waveform of (a)
22 Voltage during DOL starting at 380 V, 60 Hz, 113 Nm load torque, with a 25
coupled load inertia of 100 lb-ft2 - (a) Time waveform, (b) rms value of the
voltage waveform of (a)
23 Start with failure to synchronize – (a) Speed, (b) current 26
24 Limiting cases of synchronization capability of 286T frame prototype 27
25 Simple payback based on electricity savings with 24/7 operation 30
26 Simple payback based on electricity savings with 4000 hours/year operation 31
27 Price fluctuations of magnet constituent materials over the past four years 32
28 Soft-Start coupling used to enable synchronization of high torque or high 33
inertia loads

4
List of Tables

Table Title Page

I Performance of the tested prototype LSIPM motor at an 18
output power of 30 hp
II Performance of the tested prototype LSIPM motor at an 19
input of 380 V, 60 Hz
III Performance of the tested NEMA 440 frame prototype 21
LSIPM motor at an input of 480 V, 60 Hz
IV NEMA Efficiency Bands 22
V Starting Current and Torque Compared to NEMA Design B 23
VI Annualized US Energy Savings and CO2 Reductions 28
VII Efficiency Achieved Relative to NEMA Standards 31
VIII Patent applications submitted during this project 32

5
 1.0 Executive Summary

The primary purpose of this project was to combine the ease-of-installation and ease-of-use
attributes of industrial induction motors with the low-loss and small size and weight advantages of
PM motors to create an ultra-efficient, high power density industrial motor that can be started
across-the-line or operated from a standard, Volts/Hertz drive without the need for a rotor position
feedback device. PM motor products that are currently available are largely variable speed motors
that require a special adjustable speed drive with rotor position feedback. The reduced size and
weight helps to offset the magnet cost in order make these motors commercially viable.
The scope of this project covers horsepower ratings from 20 – 500. Prototypes were built and
tested at ratings ranging from 30 to 250 HP. Since fans, pumps and compressors make up a large
portion of industrial motor applications, the motor characteristics are tailored to those applications.
Also, since there is extensive use of adjustable frequency inverters in these applications, there is the
opportunity to design for an optimal pole number and operate at other than 60 Hz frequency when
inverters are utilized. Designs with four and eight pole configurations were prototyped as part of
this work. Four pole motors are the most commonly used configuration in induction motors today.
The results of the prototype design, fabrication, and testing were quite successful. The 50 HP
rating met all of the design goals including efficiency and power density. Tested values of motor
losses at 50 HP were 30% lower than energy efficient induction motors and the motor weight is 35%
lower than the energy efficient induction motor of the same rating. Further, when tested at the 30
HP rating that is normally built in this 286T frame size, the efficiency far exceeds the project design
goals with 30 HP efficiency levels indicating a 55% reduction in loss compared to energy efficient
motors with a motor weight that is a few percentage points lower than the energy efficient motor.
This 30 HP rating full load efficiency corresponds to a 46% reduction in loss compared to a 30 HP
NEMA Premium® efficient motor.
The cost goals were to provide a two year or shorter efficiency-based payback of a price
premium associated with the magnet cost in these motors. That goal is based on 24/7 operation with
a cost of electricity of 10 cents per kW-hr.
Similarly, the 250 HP prototype efficiency testing was quite successful. In this case, the
efficiency was maximized with a slightly less aggressive reduction in active material. The measured
full load efficiency of 97.6% represents in excess of a 50% loss reduction compared to the
equivalent NEMA Premium Efficiency induction motor. The active material weight reduction was a
respectable 14.5% figure. This larger rating demonstrated both the scalability of this technology and
also the ability to flexibly trade off power density and efficiency.
In terms of starting performance, the 30 – 50 HP prototypes were very extensively tested. The
demonstrated capability included the ability to successfully start a load with an inertia of 25 times
the motor’s own inertia while accelerating against a load torque following a fan profile at the
motor’s full nameplate power rating. This capability will provide very wide applicability of this
motor technology. The 250 HP prototype was also tested for starting characteristics, though without
a coupled inertia and load torque. As a result it was not definitively proven that the same 25 times
the motor’s own inertia could be started and synchronized successfully at 250 HP. Finite element
modeling implies that this load could be successfully started, but it has not yet been confirmed by a
test.
The conclusions reached as a result of this project include the following.
 The targeted loss reduction of 30% compared to induction motors has been demonstrated to
be achievable.

6
 The active material weight reduction target of 30% can also be achieved.
 The loss reduction and active material weight reductions are not an either/or proposition, but
rather both can be simultaneously achieved.
 The starting and synchronization performance demonstrated the ability to successfully start
loads of significant inertia.
 The operation of these motors on adjustable frequency inverters is possible without the use
of special PM control algorithms nor any shaft position feedback device.
 There are torque pulsations during starting that include torque reversals. These pulsations
are of sufficient amplitude that the selection and sizing of couplings needs to account for this
behavior.
 The commercial viability of motors using this technology is highly dependent on the cost
and availability of powerful permanent magnets.
It is recommended that motor product development should be undertaken to commercialize this
technology.

2.0 Introduction

Three phase electric motors are widely used in US industrial applications. Fans, pumps, and
compressors are three broad categories of applications operated by electric motors. The primary
type of motor used for these applications is a three-phase squirrel-cage induction motor. Induction
motors already have better efficiency than many of the devices they drive. They also are much more
energy efficient than other prime movers such as internal combustion engines. However, due to the
sheer quantity of connected power associated with electric motors, there are still very substantial
reductions possible in energy usage. This reduced energy usage then also provides an associated
reduction in CO2 emissions through the power generation process.
The goal of this project was to develop line-start and line-run constant-speed electric motors and
simple-to-control electric motors with the goal of obtaining at least a 30% reduction in motor losses
as compared to conventional energy-efficient induction motors and a 15% reduction in motor losses
as compared to NEMA Premium® efficient induction motors. These ultra-efficient motors will be
30% smaller in volume, 30% lower in weight, and have a higher power factor than energy-efficient
or NEMA Premium® induction motors, factors expected to drive rapid market penetration into user
and original equipment manufacturer markets.
The technology to enable this simultaneous improvement in both efficiency and power density is
dependent on permanent magnets. PM motor products that are currently available are largely
variable speed motors that require a special adjustable speed drive with rotor position feedback.
These motor systems are applied where rapid motor dynamic response is required and the lower
rotor inertia of the high power density PM motor is an advantage when compared to an induction
motor. However, there are many applications, such as pumps fans, and compressors, where dynamic
response requirements are very low. Pump, fan, and compressor applications utilize over 60% of
industrial electric motor energy in the US. In many of these applications constant-speed induction
motors that are started across-the-line are the motor of choice. Alternatively, variable speed
induction motors, powered from an open-loop (Volts/Hertz control) variable speed drive, are utilized
without any rotor position feedback device. Induction motor Volts/Hertz drives are commonplace
and available from a large number of drive manufacturers.
The primary objective of this project is to combine the ease-of-installation and ease-of-use
attributes of industrial induction motors with the low-loss and small size and weight advantages of

7
PM motors to create an ultra-efficient, high power density industrial motor that can be started
across-the-line or operated from a standard, Volts/Hertz drive without the need for a rotor position
feedback device. This will be accomplished by adding a starting cage to the rotor of the PM motor.
Computer simulation and design tools will be developed for these ultra-efficient motors in order to
predict the starting characteristics and to allow for design optimization. The design tools will be
verified with tests on laboratory prototypes (50 hp and 200 hp) that will be designed to meet the
requirements of project team members Colfax pump and Howden fan. Project success will be
measured by the energy efficiency and power density levels achieved and by the ability to predict
the starting and steady state performance of the prototype PM motors based on laboratory testing of
both line-start and open-loop-controllable PM motors. The open-loop-controllable PM motors will
likely require a less substantial rotor cage thereby allowing more rotor design freedom to maximum
efficiency and/or minimize motor volume and weight.
If these “hybrid” motors which combine line-starting of induction machines with excellent
efficiency of PM machines can be developed across a wide range of power ratings, the energy
savings are substantial. For an individual industrial user, the primary benefit is reduced electricity
cost. If the price premium paid for this motor can be recouped (via saved electricity) in two years or
less, that typically can drive commercial adoption of a new technology.

3.0 Background

Pumps, fans, and compressors use more than 60% of industrial electric motor energy in the
United States. The most widely used motors in these applications are constant-speed motors that are
started and run across the line. In some applications, variable-speed motors, powered from an open-
loop variable-speed drive, are utilized (without any rotor position feedback device) to achieve more
energy-efficient system operation when flow control is desirable. Induction motors are the
workhorses of industry and represent nearly the entire installed base of constant-speed and most
variable-speed motors. In the United States, induction motor efficiency for new industrial motor
sales falls into two categories: energy-efficient motors that meet the requirements of the 1992 U.S.
Energy Policy Act, and National Electrical Manufacturers Association (NEMA) Premium® efficient
motors that have even higher energy efficiency levels. NEMA Premium® motors are heavier and
have more active material than energy-efficient electric motors.
New motor technology is under development that will increase motor efficiency while reducing
the size and weight of the motor by reducing the amount of active material used. This technology
has become economically viable for some variable-speed motor applications and over a wide
horsepower range. However, its ability to be utilized in general-purpose industrial applications has
been limited by the need for a variable-speed drive with a rotor position feedback device to allow for
stable operation at any speed. This project developed the technology to create a low-loss, high
power density industrial motor that is easy to install and use and more efficient, lighter, and smaller
than current alternatives, including energy-efficient and NEMA Premium® motors. It will be a
general-purpose motor that can replace existing induction motors for a wide range of line-start and
variable-speed applications. The motor will have the ability to be started and run across the line or
operated from a standard (volts/hertz) drive without the need for a rotor position feedback device.

8
Fig 1 - Typical NEMA Premium Efficiency Induction Motor

9
 4.0 Results and Discussion

4.1 Technical approach and hypothesis guiding this approach

For this work there was a high degree of dependence on test-correlated finite-element analysis.
This is especially the case for the electromagnetic design of the rotor, including the incorporation of
both magnets and a significant starting cage.
CAGE SLOTS

Fig 2 a – Rotor configuration with radial magnetization and uniform cage slot pattern
CAGE SLOTS

Fig 2 b – Rotor configuration with azimuthal magnetization and cage pattern interrupted by magnets

10
 There is a wide range of possible magnet and cage configurations that can be used in a LSIPM
motor. Fig 1 shows two of the configurations previously proposed. In Fig 2a, the magnets are
oriented with a radial direction of magnetization. In Fig 2b, in contrast, the direction of
magnetization is azimuthal. In addition, in Fig 2a the cage slots are evenly distributed, while in Fig
2b there are “interruptions” in the cage pattern due to the magnets extending nearly to the rotor outer
diameter. The motors built and tested as part of the work reported on in this report use the V-shaped
magnet configuration as shown in Fig 3. As seen in Fig 3 the starting cage slots are distributed
around the rotor periphery – accommodating the space not claimed by the magnets. The work
reported on in this report includes rotors made with both cast and fabricated cages, with the
fabricated versions made with either aluminum or copper cage construction. The prototypes that
were the source of the majority of the test data in this project were made with cast aluminum cages
as seen in Fig 5. The rotor die casting process was as shown in Figs 6-8. The prototypes with die
cast rotors were built in a NEMA 286T frame size.
In Fig 4, a typical flux plot of an open circuit condition is shown. The pattern of flux created
by the V-shaped magnets is apparent. The V-shape provides a measure of flux concentration at
the air gap and also provides a means by which cogging torque can be minimized.
The stators used in the construction of the prototypes for testing of this technology were built
with common tooling used in induction motor manufacturing (Fig 9).

CAGE SLOTS

MAGNETS

Fig 3 – Rotor magnet and cage pattern for many of the motors tested in this project

Fig 4 – Open circuit flux pattern for the LSIPM motor of Fig. 3

11
 Balancing
sprue
Fan blade

Laminations

Fig 5 – Rotor showing end ring of die cast cage

Fig 6 – Rotor die casting manufacturing cell

12
Fig 7 – Rotor core stacked with magnets ready for die casting

Fig 8 - Shaft insertion following die casting of rotor

13
 Stator
windings

Fig 9 – Wound stator for use in prototype testing

The magnets used in these motor are high-temperature, high-energy-product, rare-earth sintered
magnets. A typical second quadrant B-H characteristic is shown in Fig 10. It can be seen that there
is a well-known temperature dependence of the magnetic properties of a NdFeB material. The
material characterized by Fig 10 can be considered a high-temperature material in that the knees of
the recoil lines stay in the third quadrant up to about 170oC. A range of magnet materials were
tested in prototype rotors, with the ability to resist demagnetization during hot starts being a key
issue. This topic is discussed further in a later section covering starting performance. There are a
number of different grades of high-energy-product magnets that can be selected based on the highest
expected temperatures

Fig 10 – Typical high-temperature rare-earth magnet second quadrant B-H curves

14
 Fig 11 – Magnetization of a prototype rotor

The magnetization of the smaller prototype rotors was done with a pulse-type magnetizer (Fig
11). For the larger, 250 HP prototype, the magnets were also magnetized with a pulse-type
magnetizer, but were done in axial rows before insertion into the rotor assembly.

4.2 Experimental methodology, test procedures, characterization methods

Note: The 30 – 50 HP prototypes are used to explain the testing methodologies employed. In
addition, due to a larger number of prototypes built in the 286T frame size, there are more test
points reported on as compared to the 250 HP 440 frame prototype.

Steady-State Performance -
The test procedure for steady-state efficiency measurements consisted of the following steps:
 The motor is first synchronized to the 60-Hz source. This can be done either by line-starting
the motor from the source or by bringing the motor up to speed using the load motor and
“paralleling” it in the fashion of a synchronous machine.
 The motor voltage and load torque are adjusted to the desired level for the test being
conducted (Fig 12a).
 Once the test is underway, the motor temperatures are monitored for the purposes of
determining when the motor has reached a steady-state operating condition. During this time
period, which is typically on the order of 3 hours or more when the motor is started from
ambient temperature, it is also often necessary to readjust the applied voltage and the load
torque to maintain them precisely at their desired values. Note that the time to reach steady
state for an interior permanent-magnet motor may be longer than that of a similarly-rated
induction motor because the temperature dependent characteristics of the permanent magnets

15
 (Fig 10) provide a feedback mechanism which can significantly impact the steady-state
operating condition. Specifically, as the motor heats up (typically due to I2R power
dissipation in the stator windings and to core loss), the magnets will heat up. This in-turn will
reduce their magnetization strength, which will in-turn cause a change in the stator current
and hence the stator-winding I2R loss which will change the magnet temperatures, etc.
 Steady-state operation is considered to have been achieved when there is no noticeable trend
in the motor temperatures. For the prototype motors, we monitor the temperatures with a
resolution of 0.5oC and have found steady-state is reached when it is clear that the
temperatures are essentially constant for a period of 30 minutes or more. Although this may
seem to be a rather stringent requirement, we have determined that it is a requirement to give
consistent, reproducible test results. This criterion must typically be tailored to the motor
under test; thermal time constants tend to increase with motor size and rating and hence the
time period over which to determine that the temperatures are constant will be
correspondingly longer for larger machines.
 Once steady-state operation has been achieved, a data acquisition system (Fig 12b) is used to
record the motor characteristics at 10 second intervals for a period of 5 minutes or more.
These include the motor terminal voltages and currents, the motor temperatures and the
electrical input power and the motor torque.
 Immediately following the recording of this steady-state data, the motor is operated open-
circuited at its rated speed and the open-circuit voltage is measured. A comparison of this
open-circuit voltage with the value measured at ambient can be used as an indirect
measurement of the magnet temperature rise.

In order to eliminate noise as well as the effects of small drifts in the motor operating condition,
the performance of the motor at any given operating conditions is determined by taking the average
of the last 5 minutes of recorded readings (30 data sets). The mechanical output power is calculated
from the product of the motor speed in rad/sec (1800 rpm = 60 rad/sec) and the average torque. The
efficiency is calculated based upon this value of output power and the average electrical input
power.

16
 Prototype
motor

Load
generator

Fig 12a – Dynamometer test setup for 30 – 50 HP steady-state tests

Power Analyzer
Oscilloscope

Thermocouple
meter

Fig 12b – Data acquisition for steady-state testing

Starting Performance -
For starting performance tests, the equipment used included both changeable inertia wheels
and an adjustable load generator as shown in Fig 13.

17
 Added
Motor Torque load Load
under test sensor inertia generator

Fig 13 – Test rig for starting and synchronization evaluation

4.3 Presentation and discussion of results

1) Measured steady-state performance of a prototype LSIPM motor at 30-hp: In this section,

typical test results are presented. Because this is a prototype motor and because one of the design
objectives is to achieve high efficiency, a series of tests at an output power of 30-hp were first
conducted to examine the impact of voltage on motor efficiency. The test results are presented in
Table I and plotted in Figs. 14-16.
From these test results, we see that it appears that this motor achieves optimal efficiency at a
terminal voltage on the order of 380 V. A second set of tests, consisting of a range of loads at 380 V,
illustrates another dimension of the performance of the tested prototype. The test results are found in
Table II and plotted in Figs. 17-19.

TABLE I
PERFORMANCE OF THE TESTED PROTOTYPE LSIPM MOTOR AT AN OUTPUT POWER OF 30 HP
Voltage Current Efficiency Power
[A] [%] factor
340 40.2 95.4 0.995
360 38.0 95.8 0.989
380 36.5 96.0 0.978
400 35.5 95.8 0.952
420 35.7 95.4 0.903
440 37.5 94.2 0.832
460 40.8 92.7 0.744

18
 TABLE II
PERFORMANCE OF THE TESTED PROTOTYPE LSIPM MOTOR AT AN INPUT OF 380 V, 60 HZ
Power Current Efficiency Power
[hp] [A] [%] factor
22.6 27.5 95.6 0.975
30.2 36.5 96.0 0.978
37.7 45.9 95.7 0.972
45.0 55.9 95.1 0.960
50.0 63.2 94.6 0.948

Fig. 14 – 30 hp tests: Efficiency vs voltage

Fig. 15 – 30 hp tests: Current vs voltage

19
Fig. 16 – 30 hp tests: Power factor vs voltage

Fig. 17 - 380 V tests: Efficiency vs power

Fig. 18 - 380 V tests: Current vs power

20
 Fig. 19 - 380 V tests: Power factor vs power

The measured steady-state performance of a larger prototype built in a NEMA 440 frame size is
shown in Table III.

TABLE III
PERFORMANCE OF THE TESTED NEMA 440 FRAME PROTOTYPE LSIPM MOTOR AT AN INPUT OF 480 V, 60 HZ
Power Current Efficiency Power
[hp] [A] [%] factor
64.9 120 95.0 0.511
125 147 96.9 0.788
192 200 97.5 0.879
250 256 97.6 0.898
279 285 97.5 0.902

The efficiency levels seen in Tables I – III for the 30 and 250 HP prototypes can be put into
context by comparing them to NEMA Premium levels. For an 1800 RPM, 30 HP Premium
Efficiency motor, the nominal efficiency is 93.6%. For a 900 RPM, 250 HP Premium Efficiency
motor, the nominal efficiency is 95.0%. Since NEMA uses a geometric progression of “efficiency
bands,” it is convenient to state changes of efficiency as a number of “efficiency band jumps.” The
table of NEMA efficiency bands is shown in Table IV. For the 30 HP prototype, the 96.0% level
represents between 5 and 6 bands higher efficiency than NEMA Premium. For the 250 HP rating,
the demonstrated 97.6% efficiency is an impressive 8 bands above NEMA Premium.

Starting Performance -
For a direct-on-line (DOL) started motor, it is necessary to demonstrate satisfactory starting
performance in addition to the steady-state performance reported above. For a LSIPM motor this
includes more than just the starting current and starting torque during the asynchronous period of
acceleration as would be the case for an induction motor. The motor also must be able to pull the
load into synchronism at the end of the starting process. Both the load torque and the load inertia
enter into whether a specific LSIPM motor will be able to successfully start and synchronize a load.
Another aspect of the starting behavior of an LSIPM motor involves the fact that the magnet flux

21
and rotor saliency introduce oscillating components of torque at all speeds until the motor is
synchronized. While there are oscillatory components of torque during DOL starting of induction
motors, rotor saliency and magnet flux create a more significant set of torque oscillations during
DOL starting of a LSIPM motor. The rotor saliency feature is created when the magnets are placed
in the interior of the rotor laminations. It results in preferred directions for magnetic fields in the
rotor. This saliency does not exist in common induction motors, but does exist in many wound-field
synchronous motors and generators.

Table IV – NEMA Efficiency Bands

22
 Figures 20 – 22 show the motor speed, phase current, and terminal voltage as measured during a
line start of the prototype motor discussed in the steady-state section of this report. For this test, the
motor was coupled to a load consisting of an inertia of 100 lb-ft2 and a programmable torque. In this
case, the torque was programmed to follow a “fan load” characteristic with the torque being
proportional to the square of the speed, reaching 113 Nm at 1800 rpm. The same 380 V condition
that was found to be optimal for the steady-state performance was used for this starting test. It can
be seen that the voltage seen at the motor terminals experienced the common sag and recovery
associated with drawing about 300 amps during the early portions of the start.
The oscillatory behavior seen in the speed signal of Fig 20 can be partly attributed to the torque
pulsations due to the combination of rotor saliency and magnet flux during asynchronous operation.
In addition, the torsional compliance of the couplings also modify this oscillatory behavior. While
the motor’s own inertia tends to filter the developed air-gap torque pulsations, there is significant
ripple in the shaft torque applied to the load. As a result, care must be taken in the selection and
application of couplings.

Since one common point of reference for starting behavior is NEMA Design B, it is instructive to
compare the starting performance of this motor on that basis. For a 30 HP 1800 RPM rating, the
prescribed minimum locked-rotor torque is 150% of rated. The start shown in Figs 20-22
demonstrated 200% torque at zero speed based on the initial rate of acceleration of the load inertia.
The reason for using the initial rate of acceleration as an indicator of the locked-rotor torque is that
the torque under truly locked conditions is a strong function of the locked-rotor angular position.
This is quite different than would be observed for an induction motor. It is an area that should be
explored in terms of industry test standards. In terms of the starting current drawn during the start
shown in Fig 21, the current exceeds the NEMA Design B maximum (for this 380 V level) by just
under 15%. This higher starting current would be equivalent to that for a NEMA Design A
induction motor. Table V summarizes the starting current and starting torque for a 30 HP prototype
compared to a 30 HP, 1800 RPM NEMA Design B motor.

Table V – Tested Prototype Data Compared to NEMA Design B Starting Current and Torque

Test (@ 380 V) NEMA Design B (@ 380 V)

Starting Current 300 A 263 A max
Starting Torque 200% 150% min

The test documented by Figs 20-22 was run using a fixed-speed alternator as a controllable power
source. The hardware used for these starting and synchronization tests is shown in Fig 13. The
oscilloscope traces of both the current and voltage were processed into the equivalent rms quantities.
The plots of Figs 21, 22 show both the instantaneous signals and these equivalent rms traces. As can
be seen in Fig 22, there is a noticeable droop in the terminal voltage due to the voltage drop across
the equivalent impedance of the source. Because starting torque is strongly affected by the motor
terminal voltage, the source impedance in a given application can significantly affect the ability of a
LSIPM motor to accelerate and synchronize a given load.

23
 2
380 V, 113 Nm fan load, 100 lbmft
2000

1500

Speed [rpm]
1000

500

0
0 2 4 6 8
Time [sec]
Fig 20 – Speed during DOL starting at 380 V, 60 Hz, 113 Nm load torque, with a coupled load
inertia of 100 lb-ft2
2
380 V, 113 Nm fan load, 100 lbmft
600

400

200
I1 Current [A]

-200

-400

-600
0 2 4 6 8
Time [sec]
(a)
2
380 V, 113 Nm fan load, 100 lbmft
500

400
filtered RMS Current [A]

300

200

100

0
0 2 4 6 8
Time [sec]
(b)

Fig 21 – Current during DOL starting at 380 V, 60 Hz, 113 Nm load torque, with a coupled load
inertia of 100 lb-ft2 - (a) Time waveform, (b) rms value of the current waveform of (a)

24
 2
380 V, 113 Nm fan load, 100 lbmft
600

400

200

Volts line-line
0

-200

-400

-600
0 1 2 3 4 5 6 7 8
Time [sec]
(a)
2
380 V, 113 Nm fan load, 100 lbmft
500

400
filtered RMS Voltage [V]

300

200

100

0
0 2 4 6 8
Time [sec]
(b)

Fig 22 – Voltage during DOL starting at 380 V, 60 Hz, 113 Nm load torque, with a coupled load
inertia of 100 lb-ft2 - (a) Time waveform, (b) rms value of the voltage waveform of (a)

In contrast to the start and successful synchronization depicted in Figs 20-22, the speed and current
measured during a test with a higher load inertia is shown in Fig 23. It can be seen in Fig 23a that
the speed does not synchronize at 1800 rpm, but rather oscillates around a speed slightly below
synchronous speed. In this case, the motor is operating asynchronously as an induction motor at an
average slip speed with speed variations produced by the torque pulsations produced by the magnet
and motor saliency as the rotor slips past the synchronous flux wave of the stator. The
electromagnetic models for the starting and synchronization performance of these motors matched
well with these test points. This demonstrates that these motors can be applied reliably.
Similarly, instead of the current settling to the nominal load current, it pulsates with a rather high
amplitude, again due to the fact that the salient rotor with its fixed magnetic excitation slips past the
stator flux wave. After about 7 seconds, when it was apparent that this case would not synchronize,
the contactor feeding the LSIPM motor was manually opened. A case such as depicted in Fig 23
would be expected to trip a motor starter. How quickly a particular starter might react to such a
situation and take the motor off line is beyond the scope of this work.

25
 2
400 V, 112 Nm fan load, 144 lbmft
2000

1500

Speed [rpm]
1000

500

0
0 2 4 6 8
Time [sec]
(a)
2
400 V, 112 Nm fan load, 144 lbmft
600

400

200
I1 Current [A]

-200

-400

-600
0 2 4 6 8
Time [sec]
(b)

Fig 23 – Start with failure to synchronize –

(a) Speed, (b) current

Another aspect of starting performance is related to the magnet exposure to potentially

demagnetizing fields during the starting process. Due to the asynchronous behavior during starting,
the magnets go through a “pole-slipping” process. This means that the high stator currents which
flow during DOL starting may expose the magnets to demagnetization depending upon the magnetic
characteristics and their operating temperature. The motor design and magnet selection must take
account of all possible operating conditions to insure that demagnetization does not occur. It is
important to consider the magnet temperature during the starting process, as the ability to resist
demagnetization (Fig 10) is significantly temperature-dependent for all viable magnet materials.
While it is true that the starting cage provides somewhat of a “shielding” effect for the magnets
in regard to transient fields, there are situations such as the final pole-slip before synchronization
that are slow events and therefore the cage provides very little shielding in that case.

26
 Using a fairly high temperature grade (33EH) magnet such as SANVAC 3230, there was no
irreversible demagnetization experienced through many tests across a range of conditions. However,
with lower temperature grades such as SANVAC 4020 or SANVAC 3625, if a direct-on-line start is
performed while the magnets are particularly hot, some demagnetization can occur.
Demagnetization is an unacceptable situation, and the selection of the magnet materials and their
proper thermal application is required for successful application. The electromagnetic models along
with the magnet manufacturer’s published data correlated well to the observed test results. This
demonstrated that the phenomenon of demagnetization is modeled adequately. While lower
temperature magnet grades are less costly, the ability to resist demagnetization is an absolute
requirement.

Fig 24 – Limiting cases of synchronization capability of 286T frame prototype

Figure 24 shows the limiting cases of successful starting and synchronization for a 50 HP
prototype tested at two different voltages with four different load inertias. Rated torque is 198 Nm,
so at 400 V, the inertia which can be synchronized at full load is about 120 lb-ft2.

5.0 Benefits Assessment

Introduction of ultra-efficient, high power density electric motors will enable US energy savings
of over 70 Trillion BTU per year and reduction in CO2 emissions of over 13 million tons per year.
Table VI shows the basis for the energy and emissions reductions. Baldor Electric Company, a
leader in the development and marketing of energy efficient electric motors for utility and industrial
applications, has developed across-the-line starting and simple-to-control permanent magnet (PM)
motor technology with the capability of obtaining at least a 30% reduction in motor losses compared

27
to conventional Energy Efficient induction motors and a 15% reduction in motor losses compared to
NEMA Premium Efficient induction motors. (Energy Efficient induction motors are those that meet
the requirements of the 1992 US Energy Policy Act). These ultra-efficient PM motors are 30%
smaller in volume, 30% lower in weight, and have higher power factor than Energy Efficient or
NEMA Premium Efficient induction motors. This will allow rapid market penetration into user and
original equipment manufacturer (OEM) markets. The Baldor project team included the leading
equipment manufacturers; Colfax Pumps, PeopleFlo, and Howden Fan and energy efficiency
conscious end-users: DuPont, Duke Energy, and Ameren Power. These companies provided
consultation in regard to the desired characteristics and application of motors built with this
technology. The across-the-line starting PM motors have been designed to replace constant speed
induction motors and the simple-to-control PM motors are designed to replace variable-speed
induction motors. Target applications will be pumps, fans, and compressors with motor ratings from
20 hp to 500 hp. These ultra-efficient, high power density PM motors can enable United States
annual energy savings of over $1.4 Billion. The smaller motor size and weight will keep the cost of
these PM motors low enough so rapid market penetration will occur in the wide range of US
industries that Baldor serves.

Table VI – Annualized Energy Savings and CO2 Emissions Reductions when 90% of the
Installed Base is Converted from Energy Efficient and Premium Efficient to LSPM Ultra
Efficient Motors
(Energy Efficient to Ultra Efficient Case)

(Premium Efficient to Ultra Efficient Case)

Source – ACE3 American Council for an Energy-Efficient Economy, June 2007 prepared by R. Neal Elliot, PhD, P.E.

PM motor products that have previously been available are largely variable speed motors that
require a special adjustable speed drive with rotor position feedback. These motor systems are
applied where rapid motor dynamic response is required and the lower rotor inertia of the high
power density PM motor is an advantage when compared to an induction motor. However, there are
many applications, such as pumps, fans, and compressors, where dynamic response requirements are
very low. Pump, fan, and compressor applications utilize over 60% of industrial electric motor
energy in the US. In many of these applications constant-speed induction motors that are started

28
across-the-line are the motor of choice. Alternatively, variable speed induction motors, powered
from an open-loop (Volts/Hertz control) variable speed drive, are utilized without any rotor position
feedback device. Induction motor Volts/Hertz drives are commonplace and available from a large
number of drive manufacturers. .

The primary objective of this project was to combine the ease-of-installation and ease-of-use
attributes of industrial induction motors with the low-loss and small size and weight advantages of
PM motors to create an ultra-efficient, high power density industrial motor that can be started
across-the-line or operated from a standard, Volts/Hertz drive without the need for a rotor position
feedback device. This was accomplished by adding a starting cage to the rotor of the PM motor.
Computer simulation and design tools were developed for these ultra-efficient motors in order to
predict the starting characteristics and to allow for design optimization. The design tools were
verified with tests on laboratory prototypes (30 - 250 hp) that were designed to meet the
requirements of project team members Colfax pump and Howden fan. Project success was
measured by the energy efficiency and power density levels achieved and by the ability to predict
the starting and steady state performance of the prototype PM motors based on laboratory testing of
both line-start and open-loop-controllable PM motors. The open-loop-controllable PM motors
require a less substantial rotor cage thereby allowing more rotor design freedom to maximum
efficiency and/or minimize motor volume and weight.

29
6.0 Commercialization

One of the primary motivations for a customer to use motors with this level of energy efficiency
is the payback based on lower electricity cost. Figure 25 below shows how gains in motor efficiency
can have a fairly quick payback depending on the motor price premium charged for the more
efficient motor. This figure is based on an electricity unit cost of 10 cents per kW-hr. It is also
based on continuous operation, as might occur on a pump in a process industry such as a refinery.
The 50 HP prototype demonstrated in this project had a “4 band jump” in efficiency compared to a
NEMA Premium level as an induction motor might have. Even with a $1500 price premium, there
could be less than a 2.5 year payback with the 10 cents/kW-hr cost of electricity. For locations such
as Hawaii and Italy where the electricity cost is much higher, the payback would be that much
quicker.

Simple Payback due to Efficiency Gains - 50 HP 1800 RPM Premium Eff base, 24/7 Operation
5
$250
4.5 $500
$1000
4
$1500 incremental price increase

3.5

3
years

2.5

1.5

0.5

0
1 1.5 2 2.5 3 3.5 4 4.5 5
Efficiency band jumps

Fig 25 – Simple payback based on electricity savings with 24/7 operation

Of course, with less than a 24/7 operation (for example 4000 hours/year), the payback would be
longer as shown in the Fig 26 below. A two shift operation with a 5 day workweek might be a
typical 4000 hours per year situation.

30
 Simple Payback due to Efficiency Gains - 50 HP 1800 RPM Premium Eff base, 4000 hours per year
10
$250
9 $500
$1000
8
$1500 incremental price increase

6
years

0
1 1.5 2 2.5 3 3.5 4 4.5 5
Efficiency band jumps

Fig 26 – Simple payback based on electricity savings with 4000 hours/year operation

7.0 Accomplishments

As a result of the success demonstrated in the development of this technology, there is a technical
viability to commercialize this technology. Whether there is an economic justification for such a
product development is dependent on the availability of permanent magnets at a reasonable cost.
During the course of this project work, the cost of magnets has fluctuated by a large amount. That
fluctuation is perhaps best generalized by looking at the commodity cost of two of the primary raw
materials that are most commonly used in the manufacture of these magnets. Figure 27 below shows
that fluctuation in the prices of Neodymium (Nd) and Dysprosium (Dy). At the magnet prices that
existed at the start of this project, the economic viability was there. At the peak of the magnet
prices, the payback would be longer than the target two year level at 10 cents per kW-hr electricity
cost.
Table VII shows the specific efficiency achieved at 30, 50, and 250 HP ratings. These levels are
compared to the NEMA Energy Efficient and NEMA Premium Efficient levels prescribed for those
ratings in a TEFC enclosure as was used for the prototypes in this project.

Table VII – Energy Efficiency Demonstrated by Test Relative to NEMA Standard Levels
Rating Tested NEMA Energy NEMA Premium
Efficient Efficient
30 HP @ 1800 RPM 96.0% 92.4% 93.6%
50 HP @ 1800 RPM 96.2% 93.0% 94.5%
250 HP @ 900 RPM 97.6% 94.5% 95.0%

31
 Fig 27 – Price fluctuations of magnet constituent materials over the past four years

Two patent applications were filed and in addition a provisional application was filed as part
of this project. Table VIII below shows the titles of these three applications. Figure 28 shows
the use of a soft-start coupling to enable synchronization of higher load inertias and torques.

Table VIII – Patent applications submitted during this project

Patent Application Title Status / Type
Rotor for a Line Start Permanent Magnet Machine Patent application
System and Method to Allow a Synchronous Motor to Patent application
Successfully Synchronize with Loads that have High Inertia
and/or High Torque
Synchronous Motor with Soft Start Element formed between the Provisional patent application
Motor Rotor and Motor Output Shaft to Successfully
Synchronize Loads that have High Inertia and/or High Torque

32
 Fig 28 – Soft-Start coupling used to enable synchronization of high torque or high inertia loads

8.0 Conclusions

The technology of LSIPM motors provides an unmatched level of energy efficiency. The ability
to improve upon and extend this technology to hundreds of horsepower has been demonstrated in
this project.
The design and modeling of these machines adds significant complexity and challenges compared
to either induction motors or inverter-fed PM motors. The more complex rotor geometry requires
careful finite element analysis in order to both achieve the maximum efficiency and also to have an
effective starting cage.

9.0 Recommendations

Based on magnet costs staying in an acceptable range, a product development to commercialize

this technology should be pursued. Baldor has initiated a “gate-based” product development process
to begin commercialization.

10.0 References

NEMA MG1- 2011

M.J. Melfi, S.D. Rogers, S. Evon, and B. Martin, “Permanent Magnet Motors for Energy
Savings in Industrial Applications,” IEEE IAS PCIC Conference Record, 2006.

M.J. Melfi, R.M. McElveen, and S. Evon, “Permanent Magnet Motors for Power Density and
Energy Savings in Industrial Applications,” IEEE IAS PPIC Conference Record, 2008.

33