See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/273759342

PWM effect on motor losses and temperature rise

Conference Paper · October 2013

CITATIONS READS

2 820

3 authors:

L. Aarniovuori Markku Niemelä

Lappeenranta – Lahti University of Technology LUT Lappeenranta – Lahti University of Technology LUT
57 PUBLICATIONS   332 CITATIONS    147 PUBLICATIONS   1,505 CITATIONS   

SEE PROFILE SEE PROFILE

J. Pyrhönen
Lappeenranta – Lahti University of Technology LUT
446 PUBLICATIONS   5,699 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Vehicle energy efficiency View project

ENTALT View project

All content following this page was uploaded by L. Aarniovuori on 26 May 2015.

The user has requested enhancement of the downloaded file.

PWM effect on motor losses and temperature rise
L.Aarniovuori, M. Niemelä and J.Pyrhönen
Lappeenranta University of Technology, Finland

Abstract
Totally enclosed fan-cooled (TEFC) 37 kW class 130 (B) temperature rise 50 Hz induction motor loss
measurement results in heat run tests with sinusoidal or PWM-supply are presented. The heat run
tests are performed with 25 Hz, 40 Hz and 50 Hz points with 92 per cent of the load using different
switching frequencies and modulation principles. Two equal-power frequency converters were used in
the test, one with vector control and another with DTC control. In addition to the heat run
measurement points, the drive efficiencies were also measured using speed – torque matrix of 16
measurement points. The motor temperature rise, losses as well as measurement accuracies in
sinusoidal supply or PWM supply are analyzed and discussed. The uncertainty analysis shows that in
the case of PWM signal, it is reasonable to analyze the uncertainty using the fundamental wave
power accuracy.

Introduction
The introduction of energy efficiency regulations around the world has made the determination of
losses and energy efficiency of the induction motor more and more important topic. Currently several
IEC standards are under revision or under development. 60034-1: “Rotating electrical machines -
Part 1: Rating and performance” [1] is under revision. In the next revision of the standard, also
converter-fed machines are expected to get their efficiency classes. Rotating electrical machines -
Part 2-1: Standard methods for determining losses and efficiency from tests (excluding machines for
traction vehicles) [2] is also under revision. The revision will have significant advances over the 1st
edition such as single preferred test method for induction motors, order of tests and measurement
points in the test are fixed and the minimum accuracy of the measurement instruments is increased.
60034-2-3: Rotating electrical machines - Part 2-3: Specific methods for determining losses and
efficiency from tests for converter-fed AC machines is under development and it is expected to be
launched during 2013 as a technical specification. The Canadian Standards Association is developing
the CSA 838: Energy Efficiency Methods for Three-Phase Variable Frequency Drive Systems. There
is also the development of Eco design standard running. It specifies the energy efficiency
requirements for complete drive modules (CDM) and power drive systems (PDS). This standard
defines the IE classes and provides limits as well as test procedures for their classification. The
voltage source converters have now been used almost during four decades, but just lately the
converter efficiency has become an object of interest due to the standardisation process and energy
efficiency level classifications.
At present, there exists an application guide IEC 60034-17: Cage induction motors when fed from
converters [3] and IEC 60034-31: Guide for the selection and application of energy-efficient motors
including variable-speed applications [4]. In [5], it is shown that different laboratories testing the same
motor came up with very different results. The Round Robin test series proved that the loss tolerance
of 15 % is reasonable. This shows that even the efficiency of grid connected machines is troublesome
to determine accurately. In case of converter-fed machines pulse-width-modulated signals create
extra challenge to the motor efficiency determination. The results of variable speed drives efficiency
measurements of three collaborating research institutes are presented in [6]-[8]. The authors suggest
that the efficiency of the VSD should be measured in an operating point matrix and presented by
contour curves. The additional harmonic losses caused by PWM methods have been studied widely in
the literature, e.g. in [9]-[15]. Still, there exists no generally accepted way to determine the additional
harmonic losses. There are, of course, several problems related to defining the efficiency of the
converter, the motor and the drive. For example, there is no standard modulation technology to be
used and there is no standard control method to be used either. The DC-link level to be used has no
standard etc. All of these have a significant impact on the efficiencies.
Heat run tests
As direct electrical and torque measurements very easily fail, heat run tests performed on electrical
machines are extremely important both for manufacturers and users when the amount of losses are
accurately defined. A manufacturer would like to have an optimally designed machine so that the
product is competitive, while a user wants to be sure that the motor temperature at full load does not
exceed the thermal limits of the insulation and thus have a negative impact on the motor lifetime.

A 37 kW industrial totally enclosed fan-cooled (TEFC) class 130 (B) temperature rise induction motor
was used in the tests. The frame size of the motor is 225. The catalogue value of the efficiency for this
motor is 93.6 % with a full load and 93.4 % with a 75 % load in a sinusoidal supply. Two equal-power
frequency converters were used in the test, one with the DTC and another with the vector control. The
vector controlled frequency converter uses symmetrical two-phase modulation, thus the switching
frequency is 2/3 of the carrier frequency.

In the temperature test measurements, the motor temperatures with a sinusoidal supply and with a
frequency converter supply were recorded. With the sinusoidal supply, the electrical input power of
the motor, the rotational speed and the shaft torque were recorded by a 500 Nm Magtrol torque
transducer TMHS 313. In the frequency converter use, the electrical input and output power of the
converter were recorded by two Yokogawa PZ4000 power analyzers. In these measurements, a 10
second record length with 1 million samples at a 1 minute interval was used. At each operating point,
the motor was running 540 minutes (nine hours) to obtain thermal equilibrium and correct slip. The
electrical powers shown here are an average value of the last 30 samples. At the end of the
measurement, the current and voltage were recorded with a 1 second time interval and a 1 s sample
time to perform the DFT analysis to obtain the fundamental voltage and current as well as to calculate
the distortion. The temperature of the machine was recorded with Fluke Hydra reading the three pre-
installed Pt-100 sensors located in the machine windings. One Pt-100 sensor was used to obtain the
ambient temperature. LabView was used to gather the data to get the same time stamp for all values.

Fluke Hydra
Labview
2 x PZ4000

U, I Frequency converter U, I
t
IM
37 kW

T, n
Line converter
Electric grid

IM
315 kW

Figure 1. Measurement setup. The motor under test is marked with a dashed line square.
The sinusoidal supply tests were carried out with 25 Hz, 40 Hz and 50 Hz frequencies. A synchronous
generator providing low THD was used to supply the motor with 25 Hz and 40 Hz sinusoidal voltages.
Both converters were driven with a frequency reference.

25 Hz operating point

The input (terminal) phase voltage of the frequency converter was set to an RMS value of 230 V with
a transformer. As a load, a DTC-controlled induction machine was used. The nominal load torque of
the 37 kW induction machine is 239 Nm. The load value was set to 220 Nm resulting in 92 per cent of
the nominal load torque. The temperature rises in Tables I, II and III are the average values of the
three Pt-100 sensors located in the stator windings. The laboratory temperature was recorded with a

2
Pt-100 sensor and its value was subtracted from the results. During the temperature tests, the
laboratory temperature varied between 25–28 C.

The numerical results of the measurements are given in Table I. The THD values provided in the
tables are calculated from the current. For the vector-controlled converter, the fundamental
component of the voltage decreases, when the switching frequency is increasing. Thus, more
fundamental wave power is needed to produce the required torque and speed. Although the
fundamental voltage is decreasing, the fundamental current component is increasing, and thus, the
voltage and current distortions decrease as the switching frequency is increased. The slip of the motor
is at its highest value when the switching frequency is low and losses are high. This shows that the
motor is behaving logically in the measurements.

Table I. Numerical results of the measurements at the 25 Hz operating point with sinusoidal
and converter supply.

Sinusoidal Vector Vector Vector DTC DTC DTC DTC

fsw [kHz] - 1 4 8 1 2 3 3.75
Ufund [p.u.] 0.477 0.499 0.498 0.497 0.515 0.514 0.514 0.513
Ifund [p.u.] 0.933 0.931 0.935 0.936 0.918 0.925 0.919 0.919
I,RMS [A] 64.47 66.01 64.70 64.74 63.92 63.95 63.50 63.60
n [rpm] 729* 731 734 733 732 733 733 734
T [Nm] 219.7 220.4 220.2 220.3 220.7 221.7 220.1 219.9
Pmech [kW] 16.66 16.87 16.93 16.91 16.94 17.02 16.92 16.90
Pout [kW] 18.30 19.18 18.70 18.71 18.82 18.79 18.62 18.60
Pin [kW] - 19.63 19.24 19.42 19.41 19.46 19.36 19.42
Pmotor,loss [kW] 1.64 2.31 1.77 1.80 1.88 1.77 1.70 1.70
THD50 [%] 1.34 21.7 2.48 2.36 1.96 1.58 1.08 1.52
trise [ C] 62.7 86.1 68.8 67.1 72.6 68.3 65.8 65.2
*supply frequency 24.84 Hz

The THD values show considerable distortion when the vector control converter is used with a 1 kHz
carrier frequency. The differences in the amount of harmonic components between 4 kHz and 8 kHz
carrier frequencies are not significant. The DTC converter produces less distorted currents than the
vector controlled converter. The temperature rise and loss rises versus sinusoidal supply are shown in
Figure 2. The DTC converter optimizes the voltage to be slightly higher than in case of the vector

25 800
T emp. (Vec)
Losses (Vec)
20 T emp.(DT C)
600
Losses (DT C)
Temperature rise [ C]

Motor loss rise [W]

15

400

10

200
5

0 0
0 1 2 3 4 5 6
Switching frequency [kHz]
controlled converter.

Figure 2. Extra temperature rise and loss changes compared with the 25 Hz sinusoidal supply
as a function of the switching frequency. The load torque is set to
92 % of the motor nominal load.
The additional temperature rise and additional motor losses produced by a frequency converter are at
their highest values when the switching frequency is low. The losses of the motor rise from 200 W up

3
to 760 W depending on the switching frequency and the converter used. The temperature rise and
loss curves do not exactly match in Figure 2. The temperature rise can be assumed to be a more
reliable indicator of the motor losses than the direct power measurement. If we assume that the
correct value of the measured torque is 220 Nm and the correct rotational speed is 732 rpm, the
losses vary by 77 W, when the torque measurement has an error of 1 Nm and 100 W if the speed has
an error of 1 rpm, respectively. The total motor losses at this operation point are around 1800 W, and
consequently, the relative loss error can be very large. The largest difference in the additional losses
produced by two different converters can be assumed to be the voltage fundamental wave amplitude.
A higher voltage means a smaller current, and therefore, dominating resistive losses will decrease.

40 Hz operating point

At the 40 Hz operating point, the frequency converters are functioning in the normal operating range,
and the motor power is 74 % of the nominal power. The sinusoidal voltage was produced by a
synchronous generator. The THD50 value of the grid voltage is 0.65 % measured at the motor
terminals. The numerical results of the measurements are given in Table II.

Table II. Numerical results of the measurements at the 40 Hz operating point with sinusoidal
and converter supply.

Sinusoidal Vector Vector Vector DTC DTC DTC DTC

fsw [kHz] - 1 4 8 1 2 3 3.75
Ufund [p.u.] 0.795 0.795 0.793 0.792 0.818 0.822 0.822 0.817
Ifund [p.u.] 0.938 0.935 0.941 0.940 0.925 0.925 0.927 0.927
I,RMS [A] 64.8 65.6 65.0 65.2 64.1 64.0 64.0 64.0
n [rpm] 1185 1183 1183 1183 1184 1184 1184 1184
T [Nm] 219.0 219.1 218.9 219.2 219.0 219.2 219.0 219.0
Pmech [kW] 27.18 27.14 27.12 27.16 27.15 27.18 27.15 27.15
Pout [kW] 29.28 29.79 29.47 29.47 29.55 29.43 29.39 29.38
Pin [kW] - 30.31 30.09 30.28 30.23 30.20 30.23 30.27
Pmotor,loss [kW] 2.1 2.65 2.35 2.31 2.4 2.25 2.24 2.23
THD50 [%] 0.80 18.03 3.33 3.38 2.77 1.68 1.30 1.74
trise [ C] 58.0 71.8 61.8 61.1 64.0 61.1 59.7 60.6

The numerical results show that the slip of the motor remains constant with all switching frequencies
for both converters. When the motor is driven with the DTC, the slip of the motor is 1 min-1 smaller
than when using the vector control. The reason for this is that the DTC converter drives the motor with
a slightly higher flux density. The current distortion shows that the DTC modulation is capable of
producing less distorted current than the symmetrical two-phase modulation. Similarly as at the 25 Hz
operating point, the DTC is driving the motor with a higher voltage than the vector control. The less
distorted, smaller current results in lower motor losses. The temperature rise values in Table III are
not in an order that was assumed. The temperature rise of the motor fed by the DTC converter with 3
kHz is smaller than the temperature rise with the 3.75 kHz switching frequency. The results above
support the results obtained at the 25 Hz operating point; the temperature rises of the motor are
smaller when the DTC converter is used instead of the vector controlled converter. The temperature
rise and loss changes compared with the 40 Hz sinusoidal supply as a function of switching frequency
are shown in Figure 3.

4
 15 750
T emp. (Vec)
Losses (Vec)
T emp.(DT C)
Losses (DT C)

Temperature rise [ C]
10 500

Motor loss rise [W]

5 250

0 0
0 1 2 3 4 5 6
Switching frequency [kHz]

Figure 3. Extra temperature rise and loss changes with different PWM supplies compared with
the 40 Hz sinusoidal supply as a function of switching frequency. The load torque is set to 92%
of the motor nominal load.
The motor losses with the sinusoidal supply are 2100 W and the temperature rise is 58.0 C. The
motor losses at the 40 Hz point increase from 120 W to 600 W and the temperature rises from 3 to 14
degrees Celsius compared to the values in sinusoidal supply. Both quantities show that the loss
increase produced by the frequency converter is slightly less than at the 25 Hz operating point. At the
40 Hz point, the motor is running cooler than at the 25 point, because the cooling of the motor is
significantly improved.

50 Hz operating point

At the 50 Hz operating point, the same load and the switching frequencies were used as at the 25 Hz
and 40 Hz points. The grid voltage with the THD50 value of 1.24 % was used instead of the generator.
The numerical results of the measurements are given in Table III.

Table III. Numerical results of the measurements at the 50 Hz operating point with sinusoidal
and converter supply.

Sinusoidal Vector Vector Vector DTC DTC DTC DTC

fsw [kHz] - 1 4 8 1 2 3 3.75
Ufund [p.u.] 0.999 0.926 0.919 0.922 0.922 0.914 0.906 0.914
Ifund [p.u.] 0.937 0.978 0.988 0.990 0.983 0.987 0.995 0.989
I,RMS [A] 64.9 68.5 68.4 68.2 68.4 68.3 68.9 68.7
n [rpm] 1484 1479 1482 1480 1479 1480 1479 1479
T [Nm] 219.7 219.5 219.5 219.9 218.9 218.6 218.6 218.2
Pmech [kW] 34.14 34.00 34.07 34.08 33.90 33.88 33.85 33.80
Pout [kW] 36.46 36.94 36.85 36.77 36.68 36.62 36.60 36.54
Pin [kW] - 37.51 37.53 37.59 37.46 37.47 37.54 37.45
Pmotor,loss [kW] 2.32 2.94 2.78 2.69 2.78 2.74 2.75 2.74
THD50 [%] 2.52 15.47 3.44 3.77 2.57 2.70 2.34 2.99
trise [ C] 59.2 71.5 66.2 63.8 67.8 66.1 66.4 65.7

At the 50 Hz operating point, the amplitude of the fundamental wave voltage of the vector- controlled
vector controlled converter is larger than the one that the DTC converter is using to drive the motor.
This is a result of different converter rectifier topologies. At the 25 Hz and 40 Hz operating points, the
situation with the fundamental wave amplitude is the opposite. The current distortion of the DTC
converter still remains smaller than the one obtained with the vector controlled converter.

5
 15 750
T emp. (Vec)
Losses (Vec)
T emp.(DT C)
Losses (DT C)

Temperature rise [ C]
10 500

Motor loss rise [W]

5 250

0 0
0 1 2 3 4 5 6
Switching frequency [kHz]

Figure 4. Extra temperature rise and loss changes compared with the 50 Hz sinusoidal supply
as a function of switching frequency. The load torque is set to 92 % of the motor nominal
torque.
The temperature rise results correspond well with the measured loss rises. The motor loss rises due
to PWM modulation is a function of the rotational speed, the modulation method and the switching
frequency.

Speed torque matrix

The 4 x 4 speed – torque matrix was measured after heating the motor to its normal operating
temperate. All energy efficiency (flux optimization) control features were turned off and the frequency
converters are driven with frequency references similarly as in the temperature rise tests. The vector
converter carrier frequency is 4 kHz and the DTC converter switching frequency is 3 kHz. First, the
efficiency in the nominal point was obtained and then in the load torque points in descending order.
The different speeds were measured also in descending order. The results are given here as iso-
efficiency contours.
Vector converter efficiency DTC converter efficiency Efficiency difference (Vector - DTC)
100 100 100
0.96
0.95

0.94

5
0.95
8

0.97

0.01
0.9

0.96
0.97

75 75 75
0.01
Torque (%)

Torque (%)

Torque (%)
0.98

0.015
0.96
0.95

0.94
0.95

0.97
0.96
0.97

50 50 50
0.9 0.
3

0.9
0.

0.9
0.9

8
01

7
0. 0. 0. 0.01
4

92
0.9

96
0.9

97
0.

95
96

25 25 25
5

25 50 75 100 25 50 75 100 25 50 75 100

Speed (%) Speed (%) Speed (%)

Figure 5. Converter efficiencies in speed – torque plane.

6
 Motor efficiency (Vector converter) Motor efficiency (DT C converter) Efficiency difference (Vector - DT C)
100 100 100

0.89
0.87 0.86

0.87
0.8.84

5
5

0.9
0.8
05
0

0.88

0.91
.0

0.89

-0.01
-0

-0.0
0.9
0.9

0.86

0.88
75 75 75 0

Torque (%)
Torque (%)

Torque (%)

05
0.9
-0
- 0.0 .00 .01

1
1 -0
0.86

0.87
5
0.89 8
0.85

0.9
0.9 0.9 0.
0.8

50 0 1 500 8 0.9 0.92 50 -0.0

.8 0.8 .86 0.8 9 1 1
0.8 0.9
0.85 0 7 0.889 0.89 0.8 8 0.9 0.91
15
0.8 4 .86 0.8 0.88 0. 5 0.8
7
0.89 0.9 -0.0
0.8 3 7 0.884 0.88 0.89
2 0.85 0.86 0.87 0.86 15
3 0.87 0.88
25 25 25 -0.0
25 50 75 100 25 50 75 100 25 50 75 100
Speed (%) Speed (%) Speed (%)

Figure 6. Motor efficiencies in speed – torque plane.

Drive efficiency (Vector) Drive efficiency (DT C) Efficiency difference (Vector - DTC)
100 100 100 0
0.86

0.84

5
0.9
0.8

0
0.

0.9
0.88

0.0
0.86
0.8

00

0.88
0.84

5
0.82

0.82

75 75 75 0
Torque (%)

Torque (%)
Torque (%)

0.0
0. 05
9 0
0.84
0.86

0
0.8

0.9
0.8

50 0 0 50 5
0. .82 .84
50 0. 0.88 00
-0.
8

0.88 82
6

8 0.86 0.88
0. 0. 0.8 0.86

0
0.0
78 0.84 0.86 7 4 5
0. 8 0.86 .00

05
0.82 0.84 76 0. 0.8 -0
25 25 8 2 0.84 25
25 50 75 100 25 50 75 100 25 50 75 100
Speed (%) Speed (%) Speed (%)

Figure 7. Drive efficiencies in speed – torque plane.

Although, the differences in temperature tests are well noticeable, when inspecting the drive
efficiencies in speed – torque plane, the drives with these two different converters can considered to
be equally efficient. The same conclusions can be drawn with both tests. In this case, the vector
controlled frequency converter has lower losses than the DTC converter, but the motor losses are
smaller when using the DTC converter. To obtain the results of the heat run tests takes several
months, but the matrix measurements can be done in the single day. When the measurement series
is planned, it has to keep in mind, what is the target of the measurements. When inspecting the
frequency converters at system level, the matrix measurement is enough accurate, but in the
frequency converter development the heat run test gives more reliable results.

Uncertainty analysis
The power measurement uncertainty in the case of PWM-voltage can be estimated using discrete
Fourier transform (DFT) and calculate the power components at each frequency and refer these to
manufacturer’s data sheet to get the uncertainties with different frequency ranges. From the control
engineering point of view, an electrical machine is a voltage-controlled current source. The losses of
the motor are dependent on the amplitudes and frequencies of the current components. The voltage
coupled to the machine terminal produces the current through the machine impedance. The different
voltage waveforms produce different current waveforms, and the best way to analyze the current is its
spectrum. Even when pulse-width-modulation is used, the fundamental wave is still transferring most
of the electric power to the motor and further via the air gap to the motor shaft. The different control
systems similarly as the same control system with different controller parameters have different
voltage and current spectra, and therefore the electric power flowing through the drive can be in
different frequency ranges. Figures 8 and 9 show the power spectrum estimates with these
converters.

The figures have been compiled so that the results of current and voltage DFTs are multiplied by each
other. The bars represent the sum of these power components with the given frequency ranges.

7
 0.8

0.7
per cent of the total electrical power
0.6

0.5

0.4

0.3

0.2

0.1

0
1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 10-11 11-12 12-13 13-14 14-15 15-16 16-17 17-18 18-19 19-20
Frequency range [kHz]

Figure 8. Vector-controlled inverter-fed induction motor electrical power as a function of

frequency. The measured results are obtained with a 37 kW motor driven with a 50 Hz
frequency reference. In the measurements, the carrier frequency is set to 4 kHz and the load
torque is 92 per cent of the nominal load. The fundamental wave frequency carries 98.0 %, and
the rest of the harmonic components under 1 kHz carry 0.26 % of the total power under 20 kHz.

0.8

0.7
per cent of the total electrical power

0.6

0.5

0.4

0.3

0.2

0.1

0
1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 10-11 11-12 12-13 13-14 14-15 15-16 16-17 17-18 18-19 19-20
Frequency range [kHz]

Figure 9. DTC inverter-fed induction motor electrical power as a function of frequency. The
measured results are obtained with a 37 kW motor driven with a 50 Hz frequency reference and
load is set to 92 per cent of the nominal torque. In the measurements, the average switching
frequency is 3 kHz and the load torque is 92 per cent of the nominal load. The fundamental
wave frequency carries 97.5 %, and the rest of the harmonic components under 1 kHz carry
0.4 % of the total power under 20 kHz. In the frequencies above 10 kHz, only 0.28 % of the total
power is flowing through the drive.

The figures show that the electrical power distribution is different with different PWM supply
conditions. The carrier frequency is clearly visible in the case of the PWM, but the switching frequency
of the DTC cannot be seen from the figure. The power analyzer accuracy is given as a sum of the
reading uncertainty and the measurement range uncertainty. The uncertainties considering the
electrical equipment used in the measurements are shown in Table IV. The power measurement
uncertainty of the power analyzers is given for sinusoidal signals with a unity power factor in the given
frequency range. Additional uncertainties arise from the leading or lagging power factors and
common-mode voltages in the measurement system. The torque transducer torque uncertainty is
below ±0.1 % of rated torque. This error includes the linearity and hysteresis errors. The minimum
speed detection of this high speed sensor is 1 rpm.

8
Table IV. Yokogawa PZ4000 power accuracy [16].

Frequency Accuracy
DC ±(0.2% of reading + 0.1% of range)
0.1 Hz f 10 Hz ±(0.2% of reading + 0.05% of range)
10 Hz f 45 Hz ±(0.2% of reading + 0.025% of range)
45 Hz f 1 kHz ±(0.1% of reading + 0.025% of range)
1 kHz f 10 kHz ±(0.15% of reading + 0.04% of range)
10 kHz f 50 kHz ±(0.3% of reading + 0.05% of range)
Power factor error
45 Hz f 66Hz + (0.15 tan )% of reading

Using the power spectrum, measurement instruments accuracies at given points and measurement
results, we can compile a measurement uncertainty matrix that gives the uncertainties of the
measurement results in each point of speed - torque matrix, Fig 10. On the left, the uncertainty of the
vector controlled frequency converter measurement is given, in the center the corresponding motor
measurement and on the right hand side the drive efficiency uncertainty. The error in input and output
power measurements of the converter or motor is considered not to correlate. The uncertainty matrix
of the DTC converter looks exactly the same as the one for the vector converter; because the overall
power measurement uncertainty is ± (0.100149% reading + 0.0253% range) for the PWM signal from
0.1 Hz to 20 kHz. Thus, the uncertainty analysis can be simplified and it is possible to use the
fundamental wave accuracy for the total active power and not making a remarkable error in the
analysis.

Converter Efficiency Uncertainty Motor Efficiency Uncertainty Drive Efficiency Uncertainty

100 0.2 100 100
0.32 4

8
0.3

0.28 0.26
0.28
0.3

0.28

0.2
0.3
0.
0.3

4
28

75 75 75
0.3 0.38

0.38
Torque (%)

Torque (%)

Torque (%)

0.36
0.34
0.32

0.3 0.3 0.28

6

2 0.3
0.3 34

0.34 0.3
0.3

0.32
0.2
0.2

0.
2

0.32
0.

24
50 0.36 0.34 50 0 0.3 50
6
8

0.36 0.34 .4 0.36 0.32

0.32
0.38 0.36 0.34
0.38

0.38 0.4 0.3

0.36
0.34

0.4 0.34 0.2

0.32

0.38 8 0.36
0.42 0.
0.42
4 0.4 0.4 2 0.38
0.36 0.2
6
0.44
0.3

0.4
25 0.44 0.42 25 4 0.38 25 8
25 50 75 100 25 50 75 100 25 50 75 100
Speed (%) Speed (%) Speed (%)

Figure 10. Uncertainties of the frequency converter, motor and drive efficiency determinations
using measurements. The lines present the uncertainty in percentage units in speed – torque
plane.
The uncertainties in the efficiencies are higher with smaller loads. In these measurements the
uncertainties are quite high. If we compare the differences in figures 5.-7 and the uncertainty figures,
we find out that the uncertainty is in all cases higher than the difference between these two drives.

Conclusions and discussion

The results here cannot be used to compare the different modulation methods or control systems
performance because the parameterizations, estimations and realizations affect the results. Although,
the heat run tests showed differences between the two converters driving the same motor. If we look
at the loss rises at default switching frequencies, 4 kHz for the vector controlled converter and 3 kHz
for the DTC converter. The motor loss rises for the vector controlled converter and the DTC converter
were 8% and 4% at the 25 Hz point, 12% and 7% at the 40 Hz point and 20% and 19% at the 50 Hz
point respectively, compared with the sinusoidal supply losses. According to this research, there is no
general rule of thumb that can be used to obtain the effect of PWM on the motor losses. The motor
temperature rises are directly proportional to the additional losses produced by the PWM in the motor.

9
The additional losses are a function of the rotational speed and the switching frequency. The iso-
efficiency contours are a valuable tool to visualize and analyze the large number of measurement
points and it is well justified when frequency converter driven systems are analyzed at system level.

Detailed analysis of electric power measurement uncertainty with PWM – signal is given. Here, the
measurement instruments can be considered as basic level accuracy devices and the measurement
instruments with similar accuracies are used in many motor labs around the world. It is well
acceptable in case of PWM signals to use power analyzer accuracy at fundamental frequency to
estimate the total uncertainty in the active power measurement. It should be kept in mind that, when
low power level and low power factor points are measured the uncertainties may be reasonably
higher than in the nominal point. The power factor error in the power measurement uncertainty
analysis must not be neglected. In this case the frequency converter efficiency uncertainties vary from
0.28 to 0.44 percent units. According to these measurements the rated point efficiency of the vector
control converter is (98.2±0.3)% and the rated point efficiency of the DTC converter efficiency is
(97.4±0.3)%. The motor efficiency at the nominal point with sinusoidal supply is 93.6%, and 91.9%
when driven with the vector controlled converter and 92.8% with the DTC converter with the same
0.3% uncertainty.

Although, the converter losses are different from each other, the total drive losses remain almost
constant. In this case, the frequency converters’ as well as the motor’s efficiency differences are
marginal. From the energy efficiency point of view, both these converters are almost equal. It should
be kept in mind that the uncertainties of the direct input – output efficiency determination with high
efficiency devices can be high. In engineering work, the temperature rise is a good indicator of motor
losses, but it is motor specific and cannot be used to compare different motors. Also, the temperature
rise gives only relative changes in the losses. The uncertainty of these measurements is greater than
the differences between the drives. When same laboratory setup with the same instruments and data
collection methods are used, the results are valid for comparison, but the uncertainty of the
measurements should be inspected if the results are compared with the results obtained by another
laboratories.

References
[1] Rotating Electrical Machines – Part 30: Efficiency Classes of Single-Speed, Three-Phase,
Cage-Induction Motors (IE-Code), Ed. 1, IEC 60034-20, Nov. 2008.

[2] Rotating Electrical Machines – Part 2-1: Standard Methods for Determining Losses and
Efficiency of Rotating Electrical Machinery From Tests (Excluding Machines for Traction
Vehicles), Ed. 1, IEC 60034-2-1, Sep. 2007.

[3] Rotating Electrical Machines – Part 17: Cage Induction motors when fed From Converters –
Application Guide, Ed. 4, IEC 60034-17, May 2006.

[4] Rotating Electrical Machines – Part 31: Guide for the Selection and Application of Energy-
Efficient Motors Including Variable-Speed Applications, Ed. 1, Draft Technical Specification,
2/1575/DTC, IEC/TS 60034-31, Sep. 2009.

[5] A. Möhle, “Determination of motor efficiency on the basis of IEC600034-2-1 Round-Robin

testing for the improvement of the standard,” in Proc. 2010 Motor Summit, Zürich,
Switzerland, Oct. 2010, pp. 38–39.

[6] K. Stockman, S. Dereyne, D. Vanhooydonck, W. Symens, J. Lemmens, W. Deprez, “Iso

efficiency contour measurement results for variable speed drives," in Proc. 19th Int Conf. on
Electrical Machines Rome, Italy, pp.1-6, Sept. 2010.

[7] D. Vanhooydonck, W. Symens, W. Deprez, J. Lemmens, K. Stockman, S Dereyne,

"Calculating Energy Consumption of Motor Systems with Varying Load using Iso Efficiency
Contours," in Proc. 19th Int Conf. on Electrical Machines Rome, Italy, pp.1-6, Sept. 2010.

10
 [8] W. Deprez, J. Lemmens, D. Vanhooydonck, W. Symens, K. Stockman, S. Dereyne, J.
Driesen, "Iso Efficiency Contours as a Concept to Characterize Variable Speed Drive
Efficiency," in Proc. 19th Int Conf. on Electrical Machines Rome, Italy, pp.1-6, Sept. 2010.

[9] E.N. Hidebrand and H. Roehrdanz, “Losses in three-phase induction machines fed by PWM
converter,” IEEE Trans. Energy Convers., vol. 16, no. 3, September 2001, pp. 228–233.

[10] A. Boglietti, A. Cavagino and A.M. Knight, “Factors affecting losses in induction motors with
non- sinusoidal supply,” in Conf. Rec. Ind. Appl. Conf. 2007, vol. 1, New Orleans, LA. pp.
1193–1199.

[11] J.-J. Lee, Y.-K. Kim, H. Nam, K.-H. Ha, J.-P. Hong, and D.-H. Hwang, “Loss distribution of
three- phase induction machines fed by pulsewidth-modulated inverter,” IEEE Trans.
Magnetics, vol. 40, no. 2, March, 2004. pp. 762–765.

[12] Y. Wu, R.A. McMahon, Y. Zhan and A.M. Knight, “Impact of PWM schemes on induction
motor losses,” in Conf. Rec 2006 IEEE 41st Industry Applications Conference, vol. 2, Oct.
2006, pp. 813–818.

[13] A. Ruderman, “Electrical machine PWM loss evaluation basics,” Energy Efficiency in Motor
Driven Systems, Heidelberg, Germany, Sep. 2005.

[14] Y. Zhan, A.M. Knight, Y. Wu and R. A. McMahon, “Investigation and comparison of inverter-
fed induction machine loss, “ in Proc. IEEE Ind. Appl. Soc. Annu. Meeting, Oct. 2008, pp. 1-6.

[15] A. Boglietti, A. Cavagnino, and A. M. Knight, “Isolating the impact of PWM Modulation on
Motor Iron Loss,” in Proc. IEEE Ind. Appl. Soc. Annual Meetings, Edmonton, Canada, Oct.
2008, pp. 1–7.

[16] Yokogawa PZ4000 user manual. Available on www.yokogawa.com

11

View publication stats