1

TESTING SYNCHROPHASOR DATA QUALITY FOR APPLICABILITY TO

TRANSMISSION SECURITY AND OPTIMISATION TOOLS
R. Lira, MSc

D. H. Wilson, PhD

K. Hay, PhD

T. Cumming

D. Cole

Psymetrix
Edinburgh, UK

Psymetrix
Edinburgh, UK

Psymetrix
Edinburgh, UK

Scottish Power
Glasgow, UK

Qualitrol
Belfast, UK

tom.cumming@
sppowersystems.com

[email protected]

[email protected]

Abstract This paper summarises a testing exercise

carried out by a utility, a Phasor Measurement Unit
(PMU) manufacturer and a developer of systems and applications to qualify the PMU for use in applications of
interest for transmission system security and optimisation.
The testing goes beyond the IEEE standard requirements
for synchrophasor measurement in terms of both the static
and dynamic requirements. It was proven that the PMU
met the requirements for the applications considered. The
features tested characterise the time-domain and frequency
domain performance and can be applied to other application areas.

The performance of PMUs has become an important

issue regarding the increasing use of WAMS for monitoring and control of power systems. A precise definition of generic accuracy requirements for PMUs is a
challenge for engineers, given the diversity of application areas. Requirements for steady state and dynamic
performance must be carefully addressed to ensure reliable performance of such applications. For oscillatory
stability applications (Figure 1), it is important that the
dynamic performance of the PMUs meets the requirements of the application.

Keywords: PMU performance, WAMS applications,

WAMS requirements, phasor accuracy, line parameter, stability monitoring.
1 INTRODUCTION
Wide Area Measurement Systems (WAMS) are increasingly used to monitor and improve the observability of power systems, in terms of static and dynamic
performance. The number of Phasor Measurement Units
(PMUs) installed around the world has been growing
rapidly. Increasingly, there is a recognition in the industry that phasor measurements and wide-area monitoring
have significant potential to improve power system
security.
This paper describes collaborative work between a
transmission system owner, a WAMS application provider and a PMU supplier to validate the PMU performance of a multifunction fault recorder*. From the utilitys perspective, standard and type testing is valuable,
but it is also necessary to have the capability to perform
tests for:
1) Applicability of the unit for the intended operational
and analytical uses of the WAMS
2) Testing of different configurations of the unit and
firmware upgrades
3) Interoperability between different models of used
PMUs within the network and from external sources
For these reasons, an in-house, repeatable test process
is required that validates the static and dynamic behaviour of the PMU for the applications. The goals of the
project included both the initial validation of the PMU
and a repeatable test process can be used for on-going
maintenance of the WAMS system.
*

The help and support received from OMICRON in this

project is acknowledged, and in particular for the provision of
a test unit for the testing exercise.

17th Power Systems Computation Conference

Figure 1: Dynamic characteristics of a system oscillation

which composes the mode shapes along the system.

The present IEEE C37.118-2005 standard for synchrophasor measurement [1] defines the steady state
performance of PMUs, but dynamic performance is not
covered. Nevertheless, different approaches for phasor
estimation and requirements under transient and dynamic conditions can be found in the literature [2, 3, 4].
Also, some steady-state requirements can be more stringent than the C37.118 requirements.
While PMU performance testing has been addressed
in literature [5, 6], it is generally not related to the specific requirements of the applications using the data. In
this paper, the testing exercise reported is related directly to the requirements of the applications, and the focus
of the testing was to qualify a specific PMU model for
use in the intended applications.
There are other sources of error, in particular because
of the accuracy of the instrumentation channels including the current and voltage transducers. It is known
those errors can be larger than the PMU itself [7]. The
instrumentation channels are important for applications
requiring accurate results, such as the line parameter
estimation. However, it is generally less significant for
dynamics application. This paper does not address instrumentation channels issues, but rather focuses on
testing and qualification of the PMU.

Stockholm Sweden - August 22-26, 2011

2
The main drivers for implementing a WAMS system
were to do with increasing the utilization of the transmission while maintaining system security. This is increasingly challenging in view of the renewable generation targets in Scotland.
The performance of the PMUs was tested for the following application requirements:
Oscillatory stability monitoring, requiring knowledge
of gain and phase, and good attenuation of aliased
components
Transient detection and reporting, requiring fast response to disturbances
Line parameter estimation, requiring high steady-state
accuracy
In this paper, the testing methodology for each of these application areas is described, and test results reported. It was concluded that the PMU was suitable for the
areas of application that were intended, within certain
boundaries, which were clearly understood through the
testing exercise.
2

APPLICATION REQUIREMENTS

2.1 Oscillatory Stability Monitoring

One of the benefits of synchrophasor monitoring over
conventional SCADA is that the dynamics of the system
are reproduced in the measurements. Since the data is
synchronized, it is possible to process the data to determine the amplitude and phase of oscillations across the
network, as well as the damping.
For monitoring a range of oscillation frequencies, it is
necessary for the PMU to have the following qualities:
Negligible attenuation within the frequency band of
interest
Negligible phase delay in the band of interest
Sufficient attenuation of aliased components to avoid
misleading results
Specifically, the performance targets for the oscillation are listed in Table 1. The values listed in Table 1
are based on requirements for a commercial phasorbased real-time oscillation monitoring application [8].
The application identifies the frequency, amplitude,
damping and relative phase of dominant oscillations
observed in the power system. The underlying measurement system should not introduce errors that could
lead to the user drawing false conclusions from the results.
The bandwidth currently used covers the electromechanical range of frequencies for local and inter-area
modes as well as governor control modes. The extended
frequency range is of interest for smaller machines, such
as distributed generators. The higher frequencies are
also of interest for control modes in generator testing.
It should be noted that the frequency band of interest
includes lower frequency inter-area modes under 1Hz
and also higher frequency local modes. The phase shift
is important for the lower frequency modes where it is
useful to compare the phase of oscillations at different
points in the network and in different signals. Phase
relationships are used in methods to identify coherent

17th Power Systems Computation Conference

oscillations and contributions to oscillations from different areas of the grid.

There can be frequencies above 10Hz present in the
power system. Sub-Synchronous Resonance (SSR), for
example, may occur between 10 and 50Hz. It is not
possible to observe higher frequency SSR issues using
phasor measurements without degrading the performance for other issues. The higher frequency components are out of the scope of measurements of PMUbased stability applications.
Frequency band of interest
In-band attenuation
Attenuation of aliasing
Phase shift in-band

0-4Hz (future 0-10Hz)

> -3dB
< -20dB
< 10 (main interest 0-1Hz)

Table 1: Target performance for oscillation monitoring.

2.2 Transient Detection and Analysis

The fast synchronized data from PMUs is very valuable for detecting and presenting wide-area disturbance
data. From a monitoring perspective, it is important to
know the precision with which the observations can be
interpreted.
The timing of the relative changes in the power system state is a useful measure for identifying the nearest
measurement to the source of the disturbance. Since a
disturbance propagates through the system with a finite
speed, the precise delay can be useful in determining a
sequence of events.
It is also important to distinguish the characteristics
of a disturbance in terms of the plant control behaviour
and step response characteristics of the measurement
process, such as overshoot and settling time. There are
few real step responses in the power system the system
does not instantaneously change state however, the
step response is a useful test because it allows comparison between devices. Figure 2 shows a standard step
response test, evaluated by:
Rise time from 10% to 90% of step
Overshoot beyond the final value
Settling time from initiating event reaching and remaining within 1% of step size
Knowing the step response of the PMU enables a
user to know the boundaries of conclusions that can be
derived from the measurement of an event. This helps to
distinguish the true behaviour of the power system from
artefacts of the measurement process. Also, if different
types of PMU are used in the same system, it is
important to know the extent to which comparisons can
be made between the types of PMU.
From the monitoring perspective, the step response is
required for information, however there may be further
specific requirements for control schemes. This would
involve not only assessment using the above criteria, but
would also involve capturing the time at which the
information was delivered by the PMU, thus identifying
the latency of the PMU itself. This is outside the scope
of this project, which focuses on monitoring rather than
control.

Stockholm Sweden - August 22-26, 2011

1.14

than the instrumentation channels. In an implementation

of the process, the errors introduced by the instrumentation channels are important and should also be considered.
In practice, more detailed accuracy requirements are
built up for a particular application of line parameter
estimation prior to operational use [7]. Different line
loading conditions and error combinations are considered. However, this example illustrates the typical values of accuracy required. Note that each line parameter
responds differently to errors in each phasor value.

Start Time
Overshoot

1.12

Settling Time
1%

Delay Time
1.1

Value (pu)

1.08
1.06

Rise Time
1.04
1.02
1
0.98
-0.1

-0.05

0.05

0.1
0.15
Time (sec)

0.2

0.25

0.3

Figure 2: Amplitude Step Response Test.

2.3 Line Parameter Estimation

With PMUs at either end of a transmission line, it is
possible to measure the line parameter in real time.
However, this application is sensitive to errors, and
measures must be taken to differentiate results with
higher confidence (i.e. the measurements that are consistent at both ends). The accuracy of the PMUs and the
instrumentation channels influence the number of coherent results obtained that can be treated with confidence.
The requirement for accurate line parameter estimation is accurate steady-state measurements. The dynamic
performance of the PMU is not relevant in this case. The
C37.118 standard requires a steady-state accuracy defined as 1% of the Total Vector Error (TVE). This error
is too great for accurate measurements of line parameter,
however PMUs can provide much better accuracy than
1% TVE during normal operating conditions.
Table 2 shows the typical measurement accuracy that
is required in order to achieve the target levels of line
parameter accuracy. The accuracy figures are taken from
a typical transmission line model, as single-end phasor
errors are introduced to examine the sensitivity to errors
in those variables. It is clear that the different line parameters will be sensitive to different measurement
errors. The following error modes are considered:
Single-end V magnitude accuracy. This is the magnitude error, in percentage, in just one end of the line,
given that the rest of measurements are correct
Single-end I magnitude accuracy. As above, for
current magnitude
Single-end V angle accuracy. This is the angle error, in degrees, in just one end of the line, given that
the rest of measurements are correct
Single-end I angle accuracy. As above, for current
angle
A line parameter measurement method should include
error-checking for data sets used, but it is clear that
better measurements lead to more consistent and accurate line parameter estimates. It should be noted that this
paper deals with the accuracy of the PMU only, rather

17th Power Systems Computation Conference

Accuracy Magnitude Error [%]

Angle Error []
Target
|V|
|I|
ang(V) ang( I )
R 3%
0.05%
8.0%
0.20
0.70
5%
0.10%
15.0%
0.32
1.10
X 3%
1.50%
6.0%
0.15
> 10
5%
3.0%
10.0%
0.25
> 10
B 3%
6.0%
2.0%
> 10
0.40
5%
10.0%
4.0%
> 10
0.65
Table 2: Typical one-end accuracy requirements for line
parameter estimation of a medium length, medium load transmission line

Another aspect of PMU testing that is important for

line parameter estimation is measurement noise. Individual short-window phasor samples tend to have higher
random measurement noise than samples using a longer
window or averaged results. It is useful to reduce measurement noise by filtering results; however this should
not filter out real variations in the power system.
3 TEST PROCEDURE
The testing process was carried out by injecting 3phase 50Hz test signals into the PMU front-end using a
relay test set that is capable of time synchronization and
COMTRADE file playback. The accuracy class of the
test set for magnitude and angle is suitable for the tests
carried out [9]. The tests can be replayed either into a
single PMU or simultaneously replayed into several
PMUs simultaneously for comparative tests. The
PMU(s) stream data to a Phasor Data Concentrator
(PDC) that captures and stores the data for later analysis. Separate tests are designed for testing specific application requirements.
3.1 Steady-State Tests
Steady-state tests involve running balanced 3-phase
steady-state signals into the PMU. The signal is maintained for a period at least of 10 seconds, and the mean
value and standard deviation are recorded relative to the
test signal. The steady-state tests include a range of
steady-state voltage and current tests in the ranges:
0.10 1.20pu voltage
0.03 1.00pu current
45 55Hz nominal frequency
45 angle step
A much smaller range of voltage and current is used
in practice for line parameter estimation. However, it is

Stockholm Sweden - August 22-26, 2011

4
necessary to configure the line parameter estimation to
provide results in value ranges where the measurements
are sufficiently accurate to yield results. Also, an assessment of the measurement accuracy is of interest for
the use of the measurements in post-event analysis.
Imbalanced test signals were applied to ensure that
there was no inter-dependence between the measurements of each phase. The steady-state tests were also run
at different frequencies to ensure that the results were
not influenced by frequency deviation from the nominal.
The phase angle tests are evaluated based on the absolute reference phase, where the 0 angle is defined at
the second transition provided by PPS clock [1]. Given
that the signal generator output was triggered by PPS,
the angle values reported by the PMU should corresponds exactly to the signal generator output.

TEST RESULTS

4.1 Steady-State Tests

It can be seen in the results in Figure 3 and Figure 4
that the magnitude error in voltage and current measurements in the range of interest for line parameter estimation is small. The voltage and current error is less
than 0.17% at 0.6pu and above.

3.2 Transient Response

The transient response is important to assess how
quickly the unit responds to a change in the system conditions. Upward and downward steps are used to check
the consistency of the delay in response. Steps are applied to both the magnitude and the angle signals.
Thus, the following transient response tests were
done:
Step V&I magnitude 1.0 1.1pu
Step V&I magnitude 1.1 1.0pu
Step V&I magnitude 1.0 0.9pu
Step V&I angle by 15 degrees
Step V&I angle by 3 degrees
The point of change of the signal is accurately defined using the synchronization of the test signal generator. The transient response is also compared with other
models of PMU to check the consistency between the
PMU under test and other PMU models.
3.3 Oscillation Analysis
The oscillation tests look at the gain and phase of oscillations through a range of frequencies. A series of
oscillation frequencies are applied with constant oscillation amplitude and known phase. The output is analysed
in terms of the gain of the oscillation and the phase
relative to the input oscillation.
The frequency responses of following signals were
tested:
Voltage and Current Amplitude
Voltage and Current Angle
Frequency
The frequencies applied include:
0.1 10Hz:
In-band response
10 20Hz:
Gain roll-off
24 26Hz:
Around Nyquist frequency
40 49Hz:
Attenuation of aliasing
Each oscillation frequency is maintained for a period
of at least 10 seconds. The magnitude and phase of the
output oscillations were compared with the input oscillations for each test.

17th Power Systems Computation Conference

Figure 3: Steady-State Voltage Amplitude Test.

Figure 4: Steady-State Current Amplitude Test.

It was noted that the voltage input is clipped above

1.35pu. The PMU has both high resolution and low
resolution (wide range) inputs that can be used simultaneously. It was found that it was necessary to use the
high resolution inputs for synchrophasor applications. It
is possible to use the low resolution inputs for disturbance recording to capture waveforms for large deviations during faults, as the same time as using the high
resolution inputs for phasor measurements. Thus, the use
of the PMU function does not restrict the units capability for disturbance capture.
The current accuracy degrades at low values because
the relative error increases when the expected value is
low. The same performance is observed for the voltage
measurement, but the applications considered do not
require high accuracy during low voltage fault on
conditions.
Figure 5 shows the cumulative error deviation from
the average for voltage and current magnitudes at 0.5
and 1.0 pu. It can be observed that the current error
presents greater deviation than voltage error. Also, the
standard deviation tends to increase considerably for
lower magnitude levels for both voltage and current
phasors.
The results reported in Table 3 were cross-checked
with other PMUs measuring the same signals, and also

Stockholm Sweden - August 22-26, 2011

Occurencies [%]

Table 3: Steady-State Voltage and Current Phase Angle

Tests.

Figure 6: Voltage error with system frequency deviation.

Figure 7: Current error with system frequency deviation.

1
VMAG
IMAG
VANG
IANG

0.9
0.8

0.5
V 0.5pu
V 1.0pu
0

Average Angle Error ()

0.172
0.158
0.186
0.180

Voltage (1.0pu)
Voltage (0.7pu)
Current (1.0pu)
Current (0.3pu)

0.005
0.01
0.015
0.02
0.025
0.03
Voltage Mag. Abs. Error Variation [%]

0.035

Normalised Standard Deviation

Occurencies [%]

at several values of angle. The results show that the

steady-state accuracy of the phase angle is better than
0.20 for voltage and current at nominal values. The
voltage phasors are not degraded at 0.7pu. The same
performance is observed for the current phasors at
0.3pu.
In terms of Total Vector Error (TVE), which the
C37.118 stipulates should be better than 1%, it may be
noted that at nominal values, the voltage phasor is accurate to 0.45% TVE and the current phasor is accurate to
0.53% TVE. These are both substantially better than the
value required by the standard. This is to be expected, as
the range of measured values of interest for line parameter estimation is smaller than the C37.118 standard defines. However, it is significant that the near-nominal
performance of the PMU is much better than the standard, and the testing process means that the line parameter application does not need to assume the limitation
that the accuracy is 1% TVE as defined by the standard.
Regarding the system frequency deviation tests,
Figures 6 and 7 show that between 46.5Hz and 55Hz,
the errors are practically independent of system
frequency. The errors degrade below 46.5Hz, however
the power system observed is not expected to survive a
frequency under 46.5Hz. In such a severe disturbance,
data on the general behaviour of the system is useful, but
high steady-state accuracy is not critical.
Unbalanced phase had no significant effect on the
steady-state accuracy. This confirms that the PMU
measurement channels are independent.
Figure 8 shows that measurement scatter can be
reduced by filtering measurement noise. The
measurement noise, as measured by the standard
deviation of the filtered signal, reduces exponentially as
the window length increases. This suggests a window
length of around 25 samples would be appropriate for
line parameter estimation, for instance. A relatively
short window is useful, enabling to capture the real
movement of the power system, without introducing
more measurement noise than necessary.

0.7
0.6
0.5
0.4
0.3
0.2
0.1

0.5

I 0.5pu
I 1.0pu
0

0.01
0.02
0.03
0.04
Current Mag. Abs. Error Variation [%]

0.05

Figure 5: Cumulative error deviation from the average for

different V&I magnitudes.

In terms of the accuracy for line parameter

estimation, it is noted that the PMU performs well for
estimation of impedance and susceptance measurements,
however
accurate
resistance
measurement
is
challenging.

17th Power Systems Computation Conference

10

20

30

40
50
60
70
Window Length (samples)

80

90

100

Figure 8: Reduction of Measurement Noise with Moving

Average Filter.

4.2 Transient Response

It is shown in Figure 9 that the V&I magnitude step
response is very fast. The PMU responds in a single
sample to the change in input. The rise time, delay time
and settling time are therefore all within one sample, i.e.
less than 20ms. There is no significant overshoot or

Stockholm Sweden - August 22-26, 2011

6
ripple. This response is as good as can be achieved with
a PMU. Similar responses were achieved for all phases,
for upward and downward steps.
The V&I step response of angle to a large angle step
(15), illustrated in Figure 10, shows a similarly short
fast rise time and delay time, which is less than one
sample. However, there is a small 1-sample ripple that
does not overshoot the target value, and the value takes
100ms to reach the 1% limit for the settling time. Similar results are obtained for voltage steps.
A smaller angle step (3, Figure 11) shows similar
performance to the 15 angle step, but with a smaller
overshoot value.

Figure 9: V&I Magnitude Step Response (0.1pu step).

4.3 Oscillation Analysis

In Figure 12, it is shown that the phasor outputs of
the PMU have attenuation above -3dB below 10Hz.
This shows that the phasor outputs will effectively reproduce frequency components in the range 0-10Hz,
which is the desired range for current and future oscillation analysis.
Also, it is shown that aliased components in the range
40-50Hz are attenuated by a factor of at least four (below -11dB). Frequencies in the range 45-50Hz are attenuated by a factor of about eight (-17dB). In practical
terms, these results meet the target value of -20dB for
the current requirements for analysis of 0-4Hz oscillations, but care should be taken in measuring higher frequency oscillations due to aliasing. The phase shift is
within 10 up to about 1.5Hz (Figure 13), as required.
As shown in Figure 14, the frequency output signal is
quite significantly attenuated in the 0-4Hz band. At 4Hz,
the attenuation is -6dB, and the -3dB point is around
2.5Hz. Also, the phase is shifted by 20 degrees at 1Hz
(Figure 15). The frequency signal does not meet the
requirements for oscillation analysis noted in Table 1.
The result indicates that it is necessary to restrict the use
of the frequency signal to a narrow bandwidth: 0-2.5Hz
if only amplitude is required or 0-0.5Hz if <10 phase
shift is required. As an alternative, a frequency signal
can be derived from the voltage phasors, as a substitute
for the frequency signal.
5
V mag
Amplitude Modulation
Phase Angle Modulation
V angle
0

V Attenuation (dB)

-5

-10

-15

-20

-25

Figure 10: V&I Angle Step Response (15 degree step).

-30

10

15

20

25

30

35

40

45

50

Input Oscillation Frequency (Hz)

Figure 12: Amplitude Response of Amplitude and Phase

Angle Modulation, 0-50Hz.

Figure 11: V&I Angle Step Response (3 degree step).

17th Power Systems Computation Conference

The phasor outputs are composed by the voltage and current phasors. In addition, the bus frequency signal is reported
by the PMU in the same data frame [1]. Frequency is generally calculated from the voltage waveform and is independent of
the phasor calculation.

Stockholm Sweden - August 22-26, 2011

100
V mag

Amplitude Modulation
V angle
Phase Angle Modulation

V Phase (deg)

50

-50

10

Input Oscillation Frequency (Hz)

Figure 13: Phase Response of Amplitude and Phase Angle

Modulation, 0-10Hz.
0

Freq. attenuation (dB)

-1

0
-5

-2

-3

-4

-10

Freq. attenuation (dB)

-5

-15
-6

-20

Input Oscillation Frequency (Hz)

-25
-30
-35
-40
-45

10

15

20

25

30

35

40

45

50

Input Oscillation Frequency (Hz)

Figure 14: Amplitude Response in Frequency Output Signal

to Frequency Modulation, 0-50Hz.
100
90
80

Freq. Phase (deg)

70
60
50
40
30
20
10
0

Input Oscillation Frequency (Hz)

Figure 15: Phase Response in Frequency Output Signal of

Frequency Modulation, 0-6Hz.

17th Power Systems Computation Conference

5 DISCUSSION
From the testing carried out, it is concluded that the
PMU under test is suitable for use in the three contrasting applications considered. In particular, it may be
noted that:
The PMU steady-state accuracy in normal situations
is significantly better than the requirement for 1%
TVE specified in the C37.118 standard. This is important for line parameter estimation.
The transient response to a change in signal level is
very fast. The rise time is short, with negligible delay
and very little settling time.
For analysing oscillations, there is very little attenuation in the band 0-10Hz for V&I measurements, however to achieve sufficient attenuation of aliased components, a bandwidth of 0-4Hz is used. This is sufficient to cover electromechanical and frequency control modes.
There is a trade-off to be made between the transient
performance of the PMU and the oscillation performance. Filtering with a sharper roll-off would improve
oscillation performance with greater attenuation of high
frequency components while providing good in-band
performance. However, such filtering would also tend to
slow the transient performance and increase the overshoot and ripple. Some further work would be of value
to investigate whether further high frequency attenuation
is possible without a significant influence on the transient
performance, so that the bandwidth 0-10Hz can be
6
used. The extended frequency range would be useful, for
example, for generator control system tuning.
The extended bandwidth could also be used if there
are no significant components of frequency in the range
40-46Hz. It is thought that such components are unusual
in the power system. Since the PMU device is also a
disturbance recorder, it would be possible to cross-check
using waveform data if observations of frequency at 410Hz were correctly represented, or were aliased components.
The testing process has shown that the V&I phasor
data is more reliable for oscillation analysis than the
frequency output. The frequency output is restricted to a
narrow bandwidth, and cannot be used directly if the
relative phase of oscillations is important. However, it is
relatively straightforward to derive a frequency signal
from voltage angle data, and this should provide a more
useful representation of dynamic frequency.
The testing process has proved to be valuable in gaining experience in configuring and using the information.
It was found, for example, that the application requirements were better served using the high resolution configuration of the PMU, rather than the option of an extended measurement range. The high resolution setting
is necessary for any PMUs used for line parameter estimation, and is strongly recommended for at least some,
if not all of the oscillation monitoring points. There may
be some locations where the extended range may be
necessary for the disturbance monitoring functions, but
where possible, the high resolution setting is preferred
for wide-area monitoring applications.

Stockholm Sweden - August 22-26, 2011

8
This paper addresses the accuracy of the PMU itself.
As mentioned previously, the instrumentation channels
are important for applications requiring accurate results,
such as line parameter estimation. The influence of the
measurement chain is reported [7] and is the subject of
further research.
The testing that has been done has focused on the use
of the PMUs as part of a monitoring system, and not for
control. Depending on the specific requirements of control applications, there may be further testing to be done
on the continuity of data and the latency between a system change and the change being reported.
6 CONCLUSIONS
The testing process has proven the capability of the
PMU to be used for three applications that serve as
examples of applications requiring:
High steady-state accuracy;
Fast response;
Accurate representation of oscillations.
It has been proven that the tested PMU satisfies the
main requirements of all three applications. This paper
relates the process and results of PMU testing to the
needs of applications, and shows how the PMU performance enables and also bounds the information that can
be derived from synchrophasors. The testing process
and results can also be applied to requirements for future
applications.
REFERENCES
[1] IEEE C37.118. "Standard for Synchrophasors for
Power Systems", 2005.
[2] Phadke, A. G.; Kasztenny, B. "Synchronized Phasor and Frequency Measurement Under Transient
Conditions". IEEE Transactions on Power Delivery.
2009.
[3] Premerlani, W.; Kasztenny, B.; Adamiak, M. "Development and Implementation of a Synchrophasor
Estimator Capable of Measurements Under Dynamic
Conditions". IEEE Transactions on Power Delivery.
2008.
[4] "PMU System Testing and Calibration Guide".
North American Synchrophasor Iniciative (NASPI),
Performance and Standart Task team. 2007.
[5] Steinhauser F. Test and Calibration of Phasor
Measurement Units, International Protection Testing Symposium, 2008.
[6] Huang, Z.; Hauer, J. F.; Martin, K. E. "Evaluation of
PMU Dynamic Performance in Both Lab Environments and under Field Operating Conditions". 2007.
[7] Lira, R., "Influence of Instrumentation Channels
Accuracy in Synchronised Phasor Measurement" (in
Portuguese), Master's Thesis, Federal University of
Santa Catarina, Brazil, 2010.
[8] Wilson D.H., Hay K., Maclaren R. F. B., Hawkins
D.J., Dunn A., Middleton A. J., Carter A., Hung W.
Control Centre Applications of Integrated WAMS-

17th Power Systems Computation Conference

Powered by TCPDF (www.tcpdf.org)

based Dynamics Monitoring and Energy Management Systems. Cigre Session, Paris, 2008
[9] OMICRON High Precision Relay test set and universal calibrator CMC 256 plus (http://
www.omicron.at /en/products/pro/secondary-testingcalibration/cmc-256plus/).

AUTHOR'S BIOGRAPHIES
Ricardo Lira received his MSc in 2010 from Federal
University of Santa Catarina, Brazil, focusing on synchrophasor measurement accuracy. He joined Psymetrix
Ltd (now part of the Alstom Group) as a Power System
Engineer in the same year. Currently working on developments in WAMS applications such as data quality and
line parameter estimation.
Dr Douglas Wilson is Chief Technology Officer for
Psymetrix Ltd. He has worked with Psymetrix since
1998, and is involved in synchrophasor measurement
systems and applications, dynamics analysis and control
and renewable generation connection. He graduated
B.Eng(Hons) and PhD from the University of Edinburgh
and MSc from the University of Manchester. He is involved in R&D, consulting and in commercial application of synchrophasor technology.
Dr Karine Hay received her PhD degree in signal
processing from University of Rennes I in 1997. She is a
senior engineer with Psymetrix Ltd, having worked in
the company since 1998. She is involved in signal processing for online power system stability measurement
and analysis.
Thomas Cumming MSc Power Systems and High
Voltage Engineering. System Monitoring Engineer for
Scottish Power, part of Iberdrola Group Responsible for
all aspects of fault recording, system monitoring, fault
analysis and location, plant condition monitoring, PD
and DGA systems. The development and application of
new technologies including PMU's and IEC61850. 35
years experience in the Power Industry.
David Cole BSc Hons - Product Manager at Qualitrol for Fault and System Disturbances responsible for
the development and subsequent sales of on-line travelling wave based fault locators and multi function disturbance recorders that include power quality and PMU.
Past experience includes testing of HV circuit breakers,
research on locating partial discharge sites in drum
lengths of polymeric cable and time with a UK Distribution company.

Stockholm Sweden - August 22-26, 2011